I'd also like to add one more thing. Yes its good to spend a lot of time learning this stuff, databases, apis, etc, but you also don't want to get stuck in what I call TUTORIAL HELL. There are so many online courses, and books, and articles telling you about a million billion ways to do the same things, and which ones are right. In my experience, learn enough that you feel you can start a project, and start working on a meaningful project, something that you'll continue working on, on like a friday night, when its cold, and nothing seems to be going right lol.
So here’s where I get confused. This sounds like it’s how to get an interview as a data scientist right? Which I understand why that would take years. There seems to be some confusion out there about what a “data analyst” does, because you’re encouraged to build “projects” for those as well which pretty much stops as visualizations that would end up on a dashboard for a sales org or something. I’m trying to upskill to get new work and I don’t have years to get there. Is there no value, employment wise, at getting good at the increments? (Database then APIs etc.?)
3:30-Example of A.P.I Twitter, Google Analytics,RUclips,Netflix,Amazon 4:06-1) How To Setup and Configure A.P.Is,2)Learn how to use libraries to help you make A.P.I calls 3)Learn How to work with data structures like JSON and Dictionaries 4:00-Use A.P.I for data sciences: 4:48-Use Cloud Database,Make Datapipeline of AWS and Google Cloud. 6:20-1)Building Models 2)Why did you pick that model can other models do the same 3)Why did you clean the data 4) What type of validation tests did you perform on the data to prepare for the model 5) Tell me about the assumptions of your model 6) How did you optimize your model 7)How did you optimize your model 8) Explain the math behind your model 7:48-Make an impact and get validation by sharing what you have done 1)Code Sharing on Data Science Subreddits or GitHub 2) Share your Insights Create Visualisations and create Blogs/ Articles or TowardsDataScience 3)Learn Django/Flask/Aws to deploy an application that can serve as an interactive dashboard or A.P.I
So helpful. I took an entire course on DA and still had no clear idea of what the process was supposed to be in practice! I also watched videos that suggested tons of projects but not exactly how to go about them. Thank you for doing this video. I feel like I finally have an inkling about what data science work entails.
There are a lot of great Data Science videos out there, but this has been the most impactful! Thanks Nate for spending a lot of time on creating these videos. Underrated channel. Maybe use ur DS skills to exploit the YT algorithm 😂
I had cleared 2 rounds for atleast 5 companies for a data scientist , some how I got left out. I feel they want something else that am not providing . Coding --> check OOPS python --> check Traditional ML Algorithms --> Check DL and NLP algo and basic working including lib's like Spacy and TensorFlow --> check Docker --> Check Will try to add all the skills you mentioned in my bucket list. Will update the results once i get something
Hi ,@@alexandrevieira3190 Was doing Masters in ML And AI in Time-Series. Just back in the race now . Have appeared in few more companies. Here are the findings , 1. Role of feature store in AWS is heavily focused or any technique which is used for selecting and storing features. 2. Shared training resources
@@kachrooabhishek I bet many of them would have been a bad fit for you anyway. Try not to think of it as rejection, think of it as an incompatibility between you and that manager
When I start studying DS I find those random intro DS videos helpful, but after gaining some insights through months you're the only one who spoke it all correctly. Just perfecting 1 pipeline is already years to master. Thanks for the video
It's kinda challenging for a DS junior who is a self-taught and is looking for internship like me to make this kind of project (because school taught me literally nothing) but it is also exciting to explore the unknown knowledge ahead. Really good video! ♥️
When i watch these videos, and lately it has been a lot watching , the more I watch, the more I read on people like you sharing their experiences on school. I have an undergrad in CSE - glad I do, it kind of helps with wanting to self-teach python, sql, and pcik up what is been shared here. My first job was in marketing (long story) so what do i know, never worked as a software developer, but many peers have said the same - a lot of the tech skills they use to pay bills now they picked up on the job.
I’m going to start my final project in the next 2 weeks. To be honest , I still don’t know which project I should build. So this helps so much in terms of planning and give me more clarity what I should focus on! Thank you so much!! 🙏✨❤️
Man, this video was just too good. There’re so many data science related videos tell ppl especially newbies how to become a data scientist. BUT this one is the most real and practical one. Thank you
What are the odds you can walk through one of your past projects step by step, showing us the entire process with all of the nit and gritty steps? It would be beyond helpful.
I have a few python videos that walk through aspects of a project but not one from start to finish. It's very time consuming but I hope in 2022, I can find some time to create a video like this
Ok thank you so much for your channel, so I’m a senior physics major right now trying to work in data science in the summer 2022, do you think I can land interviews if I have a couple upwork and/or kaggle projects on my CV?
Hey Nate, your videos are just too good. I love how your channel is so dedicated towards real word data science. By the way I noticed that you started a video series, "For your Data Science Project" and I really want you to continue making videos for this particular series because there's literally no one on RUclips with such guidance on DS projects and I have been looking for one since a very long time because I have my placements just after 12 months and I really want to make a full stack data science project. Thank you.
All ML/DS courses taught me the latter steps - data cleaning, building a model, data visualization. I learned the deployment part from different resources. So it looks like I lack the beggining steps, that is a real-life data that is updated frequently, using APIs to retrieve it and using a DB in cloud to store it. Thank you for this insightful video. I already feel inspired, I think I will try to do something with Spotify data, if it's publicly accessible :) You are a great motivator.
At first, it's daunting then after I realized that it's better to know the truth than to live believing a lie. Now I know exactly things I'm supposed to learn. Thank you so much🙏🏽
I thought about taking on an end-to-end data science project. Your insights validated my decision to bin the project I'm working on. I just wasn't see how recruiters/hiring managers would see my current project as being impactful. I.E. Locally-stored single .csv, SQL queries and maybe some pretty visuals, and a static dashboard (no streaming). Hit the nail on the head with the age of the data. I was working with 20-30 year old soccer transfer data lol Thank you for inspiring me to shift towards a full-stack project Nathan!
I have a growing interest in data science and have been posting greater attention to it. This video is a tremendous help! I was expecting to have to build 3-4 separate projects to show my skill set. Building then the way you described would help streamline the whole process. Thank you very much for putting out this video! -Bryan
Wow.... Wow.... Wow. This made me feel like I've been living in a bubble. I'm glad to get all these information from you, it's a huge step in the right direction. REAL WORLD IMPACT is the key and that's what will keep one in the right position for a job. Thanks.
I'm lucky enough that RUclips algorithm show me this really great video. I will use this as guidelines for my DS journey. Thank you so much for this great information. Keep it up !
This video left me mad inspired to dust off a project that I thought was getting too big and involved, and really market all the different ends of it. Thanks!
Finally no more Kaggle! It is indeed important to know how to retrieve and clean raw data from API cuz in reality no one’s giving you a perfect data to play with. Very insightful, love it and done subscribed!!
I've done that titanic project haha lol. but i also added up with another project on google analytics. currently doing another one on MySQL for credit risk project. never tried collecting data through api and never know how to do it. will work on that soon.
Thank you so much. Really meant it. I was at the very moment of making my decision between applying to work as an intern at a very cost due to the pandemic and doing my own DS project where I reside. Your video was a valuable input for me and made me realize that it could be something more than Kaggle projects, and I see more add-on values for me from contributing to a DS project as your described. Much appreciated!!!
I feel like this is the most helpful data science video I’ve ever seen. Now I just need to find the resources to actually learn these things lol! (Configuring APIs, making API calls, working with the twitter or Amazon API, using AWS, etc.)
I would like to express my sincere gratitude to the creator of this RUclips video for providing excellent content and being truly inspirational. The video not only taught me valuable skills in working with real data, APIs, databases in the cloud, and building models, but it also emphasized the importance of making an impact and getting validation in the field of data science. The way the creator explained the decision-making process when building models and the underlying math behind them was incredibly helpful. Additionally, the interview questions shared in the video provided valuable insights into the best practices for cleaning data, validating models, and optimizing them. Moreover, I greatly appreciate the emphasis on sharing code and insights with the data science community. Learning how to deploy an application and create interactive dashboards or APIs opens up endless possibilities for collaboration and knowledge sharing. Overall, this video has been an exceptional resource for me, and I want to thank the creator for their dedication in providing such valuable content. They have truly inspired me to continue my journey in data science and strive for excellence in every project I undertake.
Very useful Nate! I never imagined that data science could be this subjective knowing that it has the word "Science" in it. Even with your guidance, this isn't easy to implement in the real world especially if you are on your own. It seems like you have to care about an actual problem (Covid?) and then find a way somehow to apply DS technical skills to solve that said problem. It takes a surprising amount of initiative and courage to become a data scientist
Totally agree with you there. If you don’t care about the problem and are willing to share the results to get feedback, you may never actually complete the project or might just give up when you feel it’s too hard. For personal projects, I like to create things that are useful to me that might be connected to a hobby or interest. Maybe something like a classifier to help rank professional athletes and sharing that on a sports sub-reddit or analyzing metrics in a new interesting way and creating a nice interactive visualization to share. Nerd out on something you’re interesting in. That’s my advice.
This is a really good video and unintentionally made me realize I may just love this kind of thing. Thank you for making this! It's really useful and relevant.
Really good video. I guess I would reiterate the piece he said about being able to explain your work in detail as if to a child and communicate the value of the outcome. Typically I am explaining my output to children, which we also refer to as upper management, and it's often difficult to get on the same page as them. This is also mostly with charts, data and simple math analyses let alone trying to explain why machine learning is telling them something different than their gut says is right. Business value is also something to incorporate into your analyses for those working in the commercial sector. Don't just show correlations, but include the revenue impact or whatever because ultimately they are going to want to pick 3 things out of the dozen you listed to do and want the highest impact stuff. Having a great outcome story to tell is very powerful.
I have been struggling with coming up with ideas for a meaningful and not a "MNIST-type" project for the last year and found it really hard to get inspired reading different works of others. This video though made it feel so intuitive and brought me an aha-moment. Thanks for the great work Nate.
This video is amazing!! Thank you for collating all of these ideas. I have never come across your channel before, and now I am subscribed. It was refreshing to see not to worry about kaggle and get going on real life experience instead. Brilliant
As part of my EE Ms i get myself into problems of DataSince nature, tons of data from many experiments. In search to right visualization techniques, found all those fancy TimeTable aprouch which was mind blowing(hearing from you about timeStamps i was like 👉😎), and not without sorrow(many exp didnt store timeStamps, whod knew then ...). Already build some databases, didnt knew even why, just had that intuition that i should.Thank you very much for very insigtfull video, great guidence.
1. Stay away from the Titanic and Iris datasets. 2. Avoid Kaggle unless you can rank top 10. 3. Interviews are looking for data scientists with real world skills: - coding and analytics - using modern tech and tools 4. The properties of a good project: (a). Uses real data - real time data (b). Uses modern tech - APIs + Databases (c). Building models - Why this model? How did you clean your data? etc. Interviewers don't care about results as much as they care about your reasoning behind your choices. (d). Making an impact/validation - Share your code + Output your insights on graphs and make a blog or deploy an app. Make one amazing project. Then you can *almost* copy and paste your original project to have many great projects.
Hey Nate, thanks for the highly informative and to-the-point video. I was just wondering if you could also provide one or two examples of projects which fit the requirements that you mentioned? That would be great. And if you've already answered a similar question before, then kindly direct me to it, I went through quite a few comments but couldn't find a similar question which had been answered.
Depends what you're into but here's an idea...go to www.reddit.com/r/dataisbeautiful/ and try to recreate these visualizations. Try to collect the data from and API (if there is one), manipulate the data, and visualize it with whatever tool you want. Better yet, create an app that would take parameters so that the visualizations change depending on user input and serve that up on AWS or GCP. Share it on the same subreddit and see if people like it. If they don't, ask how you can improve it, and iterate from there until you build something people would want to see.
Interviewer is not your gf, you don't have to impress him or her(you are in this to enjoy because this is something you'll do for years). Just honestly share your experience, give a proper detail of what you have gone through and that's it. As an interviewer I want a candidate to share some details and don't start answering like viva exam. For those who don't have experience: You should know that market is full of fake resume and the interviewer knows that, all we expect you as a candidate to tell the truth (Yes/No and some honest details). An honest employee brings truthfulness to the work environment that later helps in taking care of technical debts. Even If you get a job by bluffing the interviewer, it'll affect you as a person in long term even in your personal life. When a candidate lies, his or her chances of learning are almost zero because they rely of manipulation more than their technical abilities.
simple. Just develop a software which buy and sell stocks automatically. And prove that your software can make money consistently and automatically every month. Your boss will sure impress. Another thing is acess data with api is common, therefore, you are acessing just common data. Learn how to scrap any data from any website even when the website require authentication and are protected from data scrapers provides more value.
The only thing about web scraping is that it's not legal and most companies won't allow their data scientists (or any one) to grab data in that way. Maybe it's overlooked for small startups but any medium size company and above would be hesitant to be scraping data from other websites (unless the site allows for it of course, but most sites do not). As for APIs, I don't mean always mean externally facing APIs. Companies have many internal APIs that can be used to grab data so knowing how to use an API would be helpful whether or not you're using an external API or internal. I like your project idea btw =)
Hi Nathan, amazing video. I was wondering if you could post links to the API's? I can't actually find them in the descriptions as you stated in the videos. Thanks again for curating and creating this relevant content. I have been subscribing to your videos and have used some of your strategies in interviews since I started my DS journey.
I liked the bit where you go over some of the model-related questions we could be asked. Maybe you could make a video that goes in more depth? I actually got asked a question like this recently at a career-fair, and I was not ready for it. I'm still pretty inexperienced.
Thank you for this great summery, very actionable and motivating, looking forward to watching your actual application of this path, even for a most simple one.
Thankyou so muchhh buddy for thissss kinda cool approach you just laid Got an absolute idea of what ima going to do next.. Again' congrats to thatt.to you Love to follow you here ❤️.
![Working with APIs in Python [For Your Data Science Project]](http://i.ytimg.com/vi/fklHBWow8vE/mqdefault.jpg)
![Natanael Cano - Amor Eterno [Official Video]](http://i.ytimg.com/vi/EO4prCs6WpM/mqdefault.jpg)
I'd also like to add one more thing. Yes its good to spend a lot of time learning this stuff, databases, apis, etc, but you also don't want to get stuck in what I call TUTORIAL HELL. There are so many online courses, and books, and articles telling you about a million billion ways to do the same things, and which ones are right. In my experience, learn enough that you feel you can start a project, and start working on a meaningful project, something that you'll continue working on, on like a friday night, when its cold, and nothing seems to be going right lol.
That's great feedback. Thank you.
Can not stress enough how right you are.
So here’s where I get confused. This sounds like it’s how to get an interview as a data scientist right? Which I understand why that would take years. There seems to be some confusion out there about what a “data analyst” does, because you’re encouraged to build “projects” for those as well which pretty much stops as visualizations that would end up on a dashboard for a sales org or something. I’m trying to upskill to get new work and I don’t have years to get there. Is there no value, employment wise, at getting good at the increments? (Database then APIs etc.?)
I M in the same loop ...n I m trying to learn wat to choose......
❤
3:30-Example of A.P.I Twitter, Google Analytics,RUclips,Netflix,Amazon
4:06-1) How To Setup and Configure A.P.Is,2)Learn how to use libraries to help you make A.P.I calls
3)Learn How to work with data structures like JSON and Dictionaries
4:00-Use A.P.I for data sciences:
4:48-Use Cloud Database,Make Datapipeline of AWS and Google Cloud.
6:20-1)Building Models 2)Why did you pick that model can other models do the same 3)Why did you clean the data 4) What type of validation tests did you perform on the data to prepare for the model
5) Tell me about the assumptions of your model 6) How did you optimize your model 7)How did you optimize your model 8) Explain the math behind your model
7:48-Make an impact and get validation by sharing what you have done 1)Code Sharing on Data Science Subreddits or GitHub 2) Share your Insights Create Visualisations and create Blogs/ Articles or TowardsDataScience 3)Learn Django/Flask/Aws to deploy an application that can serve as an interactive dashboard or A.P.I
This is probably the most honest and comprehensive advice on Internet. Fantastic video, Nate.
True
Very true
This provides so much structure. It makes it even easier for a novice to begin without the feeling of being overwhelmed because of the structure.
i can vouch, this video has definitely bringing instant perspective. thank you NAte
Wow
So helpful. I took an entire course on DA and still had no clear idea of what the process was supposed to be in practice! I also watched videos that suggested tons of projects but not exactly how to go about them. Thank you for doing this video. I feel like I finally have an inkling about what data science work entails.
Holy shit the quality of this video and amount of info in 12 mins, amazing... We need these for every work position ever
After watching this video, I will return after becoming an expert data scientist.🔥
Mindblowing. Ive watched thousands of videos on how data science projects but this one takes the throne. Subbing is a no brainer.
Thanks for the kind words!
There are a lot of great Data Science videos out there, but this has been the most impactful! Thanks Nate for spending a lot of time on creating these videos. Underrated channel. Maybe use ur DS skills to exploit the YT algorithm 😂
Thanks so much for you kind words!
The 13 minutes that I am gonna rewatch over and over again. Great advices, straight to the point.
=) Thanks!
I had cleared 2 rounds for atleast 5 companies for a data scientist , some how I got left out. I feel they want something else that am not providing .
Coding --> check
OOPS python --> check
Traditional ML Algorithms --> Check
DL and NLP algo and basic working including lib's like Spacy and TensorFlow --> check
Docker --> Check
Will try to add all the skills you mentioned in my bucket list. Will update the results once i get something
Any update?
Hi ,@@alexandrevieira3190 Was doing Masters in ML And AI in Time-Series. Just back in the race now .
Have appeared in few more companies.
Here are the findings ,
1. Role of feature store in AWS is heavily focused or any technique which is used for selecting and storing features.
2. Shared training resources
@@kachrooabhishek Keep it up! Everybody without resume experience has a 100:1 rejection ratio, but all you need is 1 first job
@@yesfredfredburger8008 thanks dude for the kind words, rejections are too dayumm painful :(
@@kachrooabhishek I bet many of them would have been a bad fit for you anyway. Try not to think of it as rejection, think of it as an incompatibility between you and that manager
When I start studying DS I find those random intro DS videos helpful, but after gaining some insights through months you're the only one who spoke it all correctly. Just perfecting 1 pipeline is already years to master. Thanks for the video
Great to get confirmation!
Wow, no more toy projects, these are real-world experiences! I appreciate that you are sharing this, great work!!!
It's kinda challenging for a DS junior who is a self-taught and is looking for internship like me to make this kind of project (because school taught me literally nothing) but it is also exciting to explore the unknown knowledge ahead. Really good video! ♥️
When i watch these videos, and lately it has been a lot watching , the more I watch, the more I read on people like you sharing their experiences on school. I have an undergrad in CSE - glad I do, it kind of helps with wanting to self-teach python, sql, and pcik up what is been shared here. My first job was in marketing (long story) so what do i know, never worked as a software developer, but many peers have said the same - a lot of the tech skills they use to pay bills now they picked up on the job.
Well i hope u have a DS job now
Hey nate, thanks for the advice, it actually feels actionable and gives me confidence!
Thanks for watching! Appreciate it
Wow. This video is the guidance I was looking for. It literally lays out the path to build the best model that a DS portfolio must have.
exactly what I was thinking
I’m going to start my final project in the next 2 weeks. To be honest , I still don’t know which project I should build. So this helps so much in terms of planning and give me more clarity what I should focus on! Thank you so much!! 🙏✨❤️
Man, this video was just too good. There’re so many data science related videos tell ppl especially newbies how to become a data scientist. BUT this one is the most real and practical one. Thank you
Thank you for the kind words! Appreciate it
This video alone is better than the whole channel of some data science youtubers.
Very insightful!
What are the odds you can walk through one of your past projects step by step, showing us the entire process with all of the nit and gritty steps? It would be beyond helpful.
I have a few python videos that walk through aspects of a project but not one from start to finish. It's very time consuming but I hope in 2022, I can find some time to create a video like this
Ok thank you so much for your channel, so I’m a senior physics major right now trying to work in data science in the summer 2022, do you think I can land interviews if I have a couple upwork and/or kaggle projects on my CV?
@@fritzahern1380 Yes, it's definitely possible
Hey Nate, your videos are just too good. I love how your channel is so dedicated towards real word data science. By the way I noticed that you started a video series, "For your Data Science Project" and I really want you to continue making videos for this particular series because there's literally no one on RUclips with such guidance on DS projects and I have been looking for one since a very long time because I have my placements just after 12 months and I really want to make a full stack data science project. Thank you.
All ML/DS courses taught me the latter steps - data cleaning, building a model, data visualization. I learned the deployment part from different resources. So it looks like I lack the beggining steps, that is a real-life data that is updated frequently, using APIs to retrieve it and using a DB in cloud to store it. Thank you for this insightful video. I already feel inspired, I think I will try to do something with Spotify data, if it's publicly accessible :) You are a great motivator.
The first video about DS projects that's actually 100% useful. Great job, Nate, I'll follow this path for my project for sure.
At first, it's daunting then after I realized that it's better to know the truth than to live believing a lie. Now I know exactly things I'm supposed to learn. Thank you so much🙏🏽
Anxiety and imposter syndrome set in. lol Thanks for the advice, I will work my way up step by step
I thought about taking on an end-to-end data science project. Your insights validated my decision to bin the project I'm working on. I just wasn't see how recruiters/hiring managers would see my current project as being impactful. I.E. Locally-stored single .csv, SQL queries and maybe some pretty visuals, and a static dashboard (no streaming). Hit the nail on the head with the age of the data. I was working with 20-30 year old soccer transfer data lol Thank you for inspiring me to shift towards a full-stack project Nathan!
I have a growing interest in data science and have been posting greater attention to it.
This video is a tremendous help! I was expecting to have to build 3-4 separate projects to show my skill set. Building then the way you described would help streamline the whole process.
Thank you very much for putting out this video!
-Bryan
One of the best Data science videos I have ever watched. Very valuable. Thank you for sharing this 💕😊
This is a gem!
Thank you! My team is happy you find our video helpful.
Very useful. Thank you!
You're welcome.
Wow.... Wow.... Wow.
This made me feel like I've been living in a bubble. I'm glad to get all these information from you, it's a huge step in the right direction. REAL WORLD IMPACT is the key and that's what will keep one in the right position for a job.
Thanks.
This video got me so excited for DS, thank you.
Thanks for watching!
I'm lucky enough that RUclips algorithm show me this really great video. I will use this as guidelines for my DS journey. Thank you so much for this great information. Keep it up !
I have been through all of these steps. I hope to land my dream job soon.
That is wonderful! Hope you land your dream job!
This video left me mad inspired to dust off a project that I thought was getting too big and involved, and really market all the different ends of it. Thanks!
Finally no more Kaggle! It is indeed important to know how to retrieve and clean raw data from API cuz in reality no one’s giving you a perfect data to play with. Very insightful, love it and done subscribed!!
I’m saving this video as a road map .lol. Thanks, I’m grateful for the info 🙌🏽
The most honest and useful video I have seen for a while about data science .. thanks
Thanks for watching!
You're the only vid on the data science community that gets to the point
I've done that titanic project haha lol. but i also added up with another project on google analytics. currently doing another one on MySQL for credit risk project. never tried collecting data through api and never know how to do it. will work on that soon.
That is awesome! Good luck with your new project.
Thank you so much. Really meant it. I was at the very moment of making my decision between applying to work as an intern at a very cost due to the pandemic and doing my own DS project where I reside. Your video was a valuable input for me and made me realize that it could be something more than Kaggle projects, and I see more add-on values for me from contributing to a DS project as your described. Much appreciated!!!
I feel like this is the most helpful data science video I’ve ever seen. Now I just need to find the resources to actually learn these things lol! (Configuring APIs, making API calls, working with the twitter or Amazon API, using AWS, etc.)
I would like to express my sincere gratitude to the creator of this RUclips video for providing excellent content and being truly inspirational. The video not only taught me valuable skills in working with real data, APIs, databases in the cloud, and building models, but it also emphasized the importance of making an impact and getting validation in the field of data science.
The way the creator explained the decision-making process when building models and the underlying math behind them was incredibly helpful. Additionally, the interview questions shared in the video provided valuable insights into the best practices for cleaning data, validating models, and optimizing them.
Moreover, I greatly appreciate the emphasis on sharing code and insights with the data science community. Learning how to deploy an application and create interactive dashboards or APIs opens up endless possibilities for collaboration and knowledge sharing.
Overall, this video has been an exceptional resource for me, and I want to thank the creator for their dedication in providing such valuable content. They have truly inspired me to continue my journey in data science and strive for excellence in every project I undertake.
ChatGPT chill please 😂
Thank you so much, the most genuine video about DS project, every concept and mistakes were explained in a very well manner.
This video needs more views honestly
Very useful Nate! I never imagined that data science could be this subjective knowing that it has the word "Science" in it. Even with your guidance, this isn't easy to implement in the real world especially if you are on your own. It seems like you have to care about an actual problem (Covid?) and then find a way somehow to apply DS technical skills to solve that said problem. It takes a surprising amount of initiative and courage to become a data scientist
Totally agree with you there. If you don’t care about the problem and are willing to share the results to get feedback, you may never actually complete the project or might just give up when you feel it’s too hard. For personal projects, I like to create things that are useful to me that might be connected to a hobby or interest. Maybe something like a classifier to help rank professional athletes and sharing that on a sports sub-reddit or analyzing metrics in a new interesting way and creating a nice interactive visualization to share. Nerd out on something you’re interesting in. That’s my advice.
@@stratascratch Thank you this is some solid advice! I feel less lost on what to do now to improve my future applications
This is a really good video and unintentionally made me realize I may just love this kind of thing. Thank you for making this! It's really useful and relevant.
Thank you for watching! We'll keep doing more!
Really good video. I guess I would reiterate the piece he said about being able to explain your work in detail as if to a child and communicate the value of the outcome. Typically I am explaining my output to children, which we also refer to as upper management, and it's often difficult to get on the same page as them. This is also mostly with charts, data and simple math analyses let alone trying to explain why machine learning is telling them something different than their gut says is right. Business value is also something to incorporate into your analyses for those working in the commercial sector. Don't just show correlations, but include the revenue impact or whatever because ultimately they are going to want to pick 3 things out of the dozen you listed to do and want the highest impact stuff. Having a great outcome story to tell is very powerful.
That's exactly right! Great write up. Agree with everything you mentioned.
This guy is spitting gold, really appreciate the depth and clarity.
Your advice is real and sharing this knowledge will help those who want to be a data scientist.
As an ambitious data scientist wannabe, I really appreciate you for this great video.
Just found this Channel and subscribed straight away, Thank you very much for spot on content and clear explanations!
thank you again.
I have been struggling with coming up with ideas for a meaningful and not a "MNIST-type" project for the last year and found it really hard to get inspired reading different works of others. This video though made it feel so intuitive and brought me an aha-moment. Thanks for the great work Nate.
Thank you for your advice. This video gives me full bandwidth of duties what a data scientist could do in real time. Thanks a lot
You are welcome!
I agree, there are so many great real-life projects you can start. Incidentally I learned Pandas on the titanic dataset ahahah.
🤣🤣
That is a popular project. Good for you :)
You are creating worldwide impact with your videos. Thank you
Thankyou! Much appreciated
Just found this and def gonna watch all you vids! You're amazing, thank you so much for sharing the valuable knowledge nate!
These information would be so useful for applying a job in DS. Thank you so much!
I love this!!! I’m a Data Science recruiter looking to jump to a more technical role! This helps a lot!!!
Thank you for such encouraging and concise video, especially for those who makes first steps in the data science path!!!
Nate, you have solid advice each time. Long-time follower and a big advocate of your channel and Stratascratch!
Thank you for your kind words and support!
@@stratascratch, well deserved!
This video is amazing!!
Thank you for collating all of these ideas.
I have never come across your channel before, and now I am subscribed.
It was refreshing to see not to worry about kaggle and get going on real life experience instead. Brilliant
This was probably the best video I’ve seen on a data science project
Glad this is helpful.
This sounds like it works for everyone, not just a data scientiests :) Thank you for the video )
Realistic! Great! Thank you for sharing your professional insight.
Thanks for watching! Seems like people like career advice more than watching me code!
I am from Taiwn, and thank you so much.
Welcome!
As part of my EE Ms i get myself into problems of DataSince nature, tons of data from many experiments. In search to right visualization techniques, found all those fancy TimeTable aprouch which was mind blowing(hearing from you about timeStamps i was like 👉😎), and not without sorrow(many exp didnt store timeStamps, whod knew then ...). Already build some databases, didnt knew even why, just had that intuition that i should.Thank you very much for very insigtfull video, great guidence.
I'm glad you got some validation from your past work! =)
Great video! I saw it yesterday and today I'm starting my project! I'll come back soon to tell you I got my job!
That is wonderful! Good luck on that job.
Great video. Provided insight for my MSc thesis.
The most practical video. Period.
Ah that is great to hear. Thank you.
Brother, thank you so much. Big help. Take Love.
1. Stay away from the Titanic and Iris datasets.
2. Avoid Kaggle unless you can rank top 10.
3. Interviews are looking for data scientists with real world skills:
- coding and analytics
- using modern tech and tools
4. The properties of a good project:
(a). Uses real data - real time data
(b). Uses modern tech - APIs + Databases
(c). Building models - Why this model? How did you clean your data? etc. Interviewers don't care about results as much as they care about your reasoning behind your choices.
(d). Making an impact/validation - Share your code + Output your insights on graphs and make a blog or deploy an app.
Make one amazing project. Then you can *almost* copy and paste your original project to have many great projects.
Great summary!
Thanks for advices!
You bet!
Hey Nate, thanks for the highly informative and to-the-point video. I was just wondering if you could also provide one or two examples of projects which fit the requirements that you mentioned? That would be great. And if you've already answered a similar question before, then kindly direct me to it, I went through quite a few comments but couldn't find a similar question which had been answered.
Depends what you're into but here's an idea...go to www.reddit.com/r/dataisbeautiful/ and try to recreate these visualizations. Try to collect the data from and API (if there is one), manipulate the data, and visualize it with whatever tool you want. Better yet, create an app that would take parameters so that the visualizations change depending on user input and serve that up on AWS or GCP. Share it on the same subreddit and see if people like it. If they don't, ask how you can improve it, and iterate from there until you build something people would want to see.
@@stratascratch Great, will have a look. Thanks for the reply!
Really good input to start with. Thank you for your time.
Interviewer is not your gf, you don't have to impress him or her(you are in this to enjoy because this is something you'll do for years). Just honestly share your experience, give a proper detail of what you have gone through and that's it. As an interviewer I want a candidate to share some details and don't start answering like viva exam.
For those who don't have experience: You should know that market is full of fake resume and the interviewer knows that, all we expect you as a candidate to tell the truth (Yes/No and some honest details). An honest employee brings truthfulness to the work environment that later helps in taking care of technical debts. Even If you get a job by bluffing the interviewer, it'll affect you as a person in long term even in your personal life. When a candidate lies, his or her chances of learning are almost zero because they rely of manipulation more than their technical abilities.
simple. Just develop a software which buy and sell stocks automatically. And prove that your software can make money consistently and automatically every month. Your boss will sure impress. Another thing is acess data with api is common, therefore, you are acessing just common data. Learn how to scrap any data from any website even when the website require authentication and are protected from data scrapers provides more value.
The only thing about web scraping is that it's not legal and most companies won't allow their data scientists (or any one) to grab data in that way. Maybe it's overlooked for small startups but any medium size company and above would be hesitant to be scraping data from other websites (unless the site allows for it of course, but most sites do not). As for APIs, I don't mean always mean externally facing APIs. Companies have many internal APIs that can be used to grab data so knowing how to use an API would be helpful whether or not you're using an external API or internal.
I like your project idea btw =)
nice advise Nath and sharing your tips to avoid unnecessary steps.
Thanks for watching as always!
Hi Nathan, amazing video. I was wondering if you could post links to the API's? I can't actually find them in the descriptions as you stated in the videos. Thanks again for curating and creating this relevant content. I have been subscribing to your videos and have used some of your strategies in interviews since I started my DS journey.
Just followed like 2 minutes into this video since I realized you are ACTUALLY about useful content and not more bs
This was a good watch, this video was the guidance I needed
Excellent!! You ignite a spark of idea of how I should go about my project. Thank you, sir.
That is just wonderful to hear. I am excited for you and your project.
I liked the bit where you go over some of the model-related questions we could be asked. Maybe you could make a video that goes in more depth? I actually got asked a question like this recently at a career-fair, and I was not ready for it. I'm still pretty inexperienced.
Thank You Nate for such an inspirational beautiful video. Thank you so much..
Couldn't help but liked and subscribed straight away. Cheers Nate
Thanks so much for subscribing!
Great breakdown, thanks! I intuitively felt that using datasets like Titanic are not very valuable, so always used my own data for learning.
Thanks for such a comprehensive set of advice. I truly agree with it.
You're welcome! I appreciate your comment.
Thank you for this great summery, very actionable and motivating, looking forward to watching your actual application of this path, even for a most simple one.
Damn this video is way too good. Better than my Christmas gift. Thank you!
Beautiful Video. Thank you
This video is pure gold.
Excellent video! Thanks so much for this! Really helps sift out irrelevant skills from relevant ones!
Thanks so much for sharing.
Thankyou so muchhh buddy for thissss kinda cool approach you just laid
Got an absolute idea of what ima going to do next..
Again' congrats to thatt.to you
Love to follow you here ❤️.
Thanks for sharing valuable info about real world Data science project needs.
Dude wow. AMAZING video. I love your channel, PLEASE keep doing this
Thanks for watching!
Ill come back to say thank you on jjob offers, after following the framework laid out in this video
This is a life changing video, thank you!
That's an excellent explanation how it can be done, thank you friend ❤️
Thanks for sharing
You're welcome. We hoped our video helped you.