Perform EDA In Seconds With Visualization Using SweetViz Library

Поделиться
HTML-код
  • Опубликовано: 5 янв 2025

Комментарии • 90

  • @saikatchakraborty1937
    @saikatchakraborty1937 4 года назад +14

    This is a very nice sir :). Like this, another library autoViz is also a very wonderful library that gives various plotting for single huge datasets within a single line of code. Thank you !!

  • @82mast
    @82mast 4 года назад +7

    wow....these detailed insights graph on hover over..i used to create in Tableau. Thanks for introducing to this new EDA library krish.

  • @RK_97
    @RK_97 4 года назад

    Thanks for the video on this new package. Big shout out for the quality of your video, comparing the older videos, this video and audio quality is great.
    I am no one to comment on your skill and the contribution that you are doing.

  • @Arjun147gtk
    @Arjun147gtk 4 года назад +2

    Wow. This is amazing. It will reduce a lot of work.

  • @annazhu3340
    @annazhu3340 2 года назад

    AMAZING library! Really appreciate your tutorial!

  • @thezeroquotient
    @thezeroquotient 4 года назад +1

    Glad I discovered this video! Great content Krish Naik!

  • @saipavan5194
    @saipavan5194 4 года назад +1

    Wow just wow and wonderful amazing visualization with simple code and I appreciate you sir for such hard dedication to teach us.

  • @raghavsharma6430
    @raghavsharma6430 4 года назад

    yes krish sir this is amazing library and thank u for keep guiding us with this sort of stuff.

  • @madhavanrangarajan6097
    @madhavanrangarajan6097 4 года назад

    This is next level sir! :) ..wow!..amazed by now!

  • @salmimabrouka4562
    @salmimabrouka4562 4 года назад

    Thank you krish for introducing this library, it's really helpful

  • @sanjogh777
    @sanjogh777 4 года назад +1

    That's pretty cool. I am gonna use it. Thanks man.

  • @natarajanlalgudi
    @natarajanlalgudi 4 года назад

    Great job Krish as always...

  • @avirupsarkar5232
    @avirupsarkar5232 4 года назад

    Superb sir.. Thank you. Keep making such awesome videos..

  • @DatascienceConcepts
    @DatascienceConcepts 4 года назад

    Really like your content! Keep up the good work.

  • @azizulhakim1534
    @azizulhakim1534 4 года назад

    just amazing !!
    thnx a lot Krish Sir.

  • @vandanpandya9587
    @vandanpandya9587 4 года назад

    thank you sir for making us familiar with such wonderful libraries

  • @AshishKumar-xx3dx
    @AshishKumar-xx3dx 4 года назад

    Thank you for showing this simple way to do EDA

  • @SyedaAbbas-x7d
    @SyedaAbbas-x7d 9 месяцев назад

    Amazing Job Sir, thanks

  • @darshitsolanki7352
    @darshitsolanki7352 4 года назад

    Amazing krish 🤘 i ❤ed ur video really helped me for my python skills

  • @samtells7702
    @samtells7702 4 года назад

    Nice video sir..
    When my target was on categorical column it didn't analyse saying, for now it can work on numerical values and boolean values

  • @jadhavnandunaik7870
    @jadhavnandunaik7870 3 года назад

    Thank you so much, for sharing your knowledge with us Sir...

  • @sairamsubramaniam8316
    @sairamsubramaniam8316 4 года назад

    Awesome! Please do a video on automl and teapot

  • @chandanmalla
    @chandanmalla 4 года назад

    EDA is not only about creating beautiful graphs, EDA helps you to understand whether the existing features are useful or not and lot of other things, try to use seaborn and matplpot lib only. And if you analyse NLP with sweetwiz I am sure your system will hang.

  • @mashakpatel4962
    @mashakpatel4962 4 года назад +1

    Sir please make video on data scientist job in finance domain and skills required for this field, starting salary and salary after five years in this domain..

  • @arjyabasu1311
    @arjyabasu1311 4 года назад

    Amazing video sir !!!!

  • @surajkhanna1129
    @surajkhanna1129 4 года назад

    Thank you so much sir for the video...explained very well!

  • @kanhaiyakumarsingh4806
    @kanhaiyakumarsingh4806 4 года назад +1

    The report didn't pop when done with Kaggle Notebook and was not saved also anywhere

  • @venud1800
    @venud1800 4 года назад +1

    Please continue the Docker Series

  • @sahanjayawarna4894
    @sahanjayawarna4894 4 года назад

    Wow, this is very efficient...

  • @sandipansarkar9211
    @sandipansarkar9211 4 года назад

    Great explanation. Need to get my hands dirty in jupyter notebook. Thanks

  • @SudipPandey
    @SudipPandey 4 года назад

    Very nice sir..

  • @kunalarora1740
    @kunalarora1740 4 года назад

    There is another very strong EDA library dataprep.eda that i have recently gone through...I personally feel it is probably faster and much more reliable than pandas profiling

    • @krishnaik06
      @krishnaik06  4 года назад +2

      Let me explore

    • @kunalarora1740
      @kunalarora1740 4 года назад

      There is a very good blog on towards datascience... You can explore through it

    • @sankhabanerjee3533
      @sankhabanerjee3533 4 года назад

      “Dataprep.eda: Accelerate your EDA” by Slavvy Coelho link.medium.com/YvLBB4Rm66
      Here's the link for the blog of dataprep.eda
      Sir you can watch it out once.
      Surely powerful than Pandas Profiling

  • @rajkumarn7201
    @rajkumarn7201 4 года назад +1

    Thanks, really helpful 👍
    Can you make a video regarding python codes for all kinds of hypothesis testing used in industrial projects, because while searching internet, I am getting chi sq, ANOVA, basic stuffs.. But in interview they are asking some testing which I am not heard off. So walking thru the overview of each hypothesis testing (if possible python code) will be really helpful

    • @adamsmohammed4499
      @adamsmohammed4499 4 года назад

      R programing language has fantastic libraries for statistics and statistical test. if you that too you can try out stat models in python which actually a package.

    • @rajkumarn7201
      @rajkumarn7201 4 года назад

      @@adamsmohammed4499 ok, but my requirement is about knowing what all statistical tests are there; just providing the list of all statistical tests used in industries will be a great help, so we can use internet to know deeper

  • @subasgiri5364
    @subasgiri5364 4 года назад

    I tried running it for a categorical target variable but it is showing an error. TARGET values can only be of NUMERICAL or BOOLEAN type for now.
    CATEGORICAL type was detected. so I think this will work only for the dataset with regression problems

  • @AI-in-Action952
    @AI-in-Action952 3 года назад

    Hi Krish , how to download the sweetviz out and render it as html in web app

  • @narendra_nn
    @narendra_nn 4 года назад

    Can we use this on datasets with no target feature..... As there is a parameter in which target feature should be specified.....?

  • @kirankhanna7829
    @kirankhanna7829 3 года назад

    How to remove Sweetviz logo from report. I am using following
    sv.config_parser.read("Override.ini")
    show_logo = 0
    Pls help

  • @arrow_the_keralite.1433
    @arrow_the_keralite.1433 4 года назад

    Thank you sir .

  • @anupammitra
    @anupammitra 4 года назад

    Is this library being used in enterprise. What are the prerequisites before using this library

  • @mahipalreddydaggula2183
    @mahipalreddydaggula2183 4 года назад

    Sweetviz is not working if the target variable is categorical.why?

  • @deepankarmullick3121
    @deepankarmullick3121 4 года назад

    its amazing

  • @True_Feelingsss...
    @True_Feelingsss... 4 года назад

    Sir please make video on missingno for missing values

  • @juniordataengineer
    @juniordataengineer 4 года назад

    Hello, Can you display the visualisations inside collab?

  • @sandeepbheri3577
    @sandeepbheri3577 4 года назад

    While importing sweetviz I am getting the following error:
    AttributeError: module 'sweetviz' has no attribute 'from_dython'
    Any work around for this ?

  • @deepakumari315
    @deepakumari315 4 года назад

    can we just use this function sweetviz.analyze in python 3.8?

  • @harish00784
    @harish00784 3 года назад

    What if i have 5 lakh records? Can I still try it out with sweetviz

  • @hadiali5922
    @hadiali5922 4 года назад

    Are these target variables taken by default??
    If so, on what basis does it chooses the target variables?

  • @sumithrap3155
    @sumithrap3155 4 года назад

    How can I see all plot on the Jupiter notebook itself

  • @anilkrajamoni1484
    @anilkrajamoni1484 4 года назад

    Pretty cool .. reduce a lot of work. :D

  • @darpan810
    @darpan810 4 года назад +3

    No such keys(s): 'compute.use_numexpr'" getting this error while using

    • @anonymous-jw3ow
      @anonymous-jw3ow 4 года назад +2

      OptionError: "No such keys(s): 'compute.use_numexpr'"
      Am also getting the same error!!!!

  • @tusharrewatkar3805
    @tusharrewatkar3805 4 года назад

    All things are being automated,then what will be the difference maker sir?

  • @SudipPandey
    @SudipPandey 4 года назад

    Hello sir,will realy appreciate if you can upload videos on so that it will be easy to understand how models are deployed and what is the coding required inside the program..I have been trying to find this on internet for many days and there is no direct material which address this..

    • @krishnaik06
      @krishnaik06  4 года назад

      Check my deployments playlist

  • @yogenderkushwaha5523
    @yogenderkushwaha5523 4 года назад

    In my case, by hovering over them the further diagrams are not visible

  • @sachinjoshi187
    @sachinjoshi187 4 года назад

    Hello Krish, I used the Pandas Profiling library for 69,000 recorda and it worked absolutely fine however same data set I tried with SweetViz and it gave me error. error " Column xyz has mixed inferred_type as determined by Pa das. This is currently not supported , column type should not contain mixed data e.g. Only Float or Str8ngs , bit not a combination". This means Pandas Profiling can work with mixed data but Swwetviz can not. However I really love your videoa and learning alot . Thanks so much Sir.

    • @krishnaik06
      @krishnaik06  4 года назад

      Just try with the same dataset with pandas profiling..I got some memory issues

  • @balavivek123
    @balavivek123 3 года назад

    Does this library eats up the memory ? I have executed as you showed but the process is never ending :)

  • @AJaditya123
    @AJaditya123 4 года назад

    Hi Krish, can we use this EDA for unsupervised learning? I see you have given target variable which is a predictable value in COMPARE fucntion.

    • @rishabhsingh8416
      @rishabhsingh8416 2 года назад

      yes, we can, you don't have to define a target variable.it is optional

  • @jayasreemangalagiri971
    @jayasreemangalagiri971 4 года назад

    Thank you krish for this demo, I have a question though...How do we share this report?

  • @littlerashi4606
    @littlerashi4606 4 года назад

    Is it not working in colab?

  • @galymzhankenesbekov2924
    @galymzhankenesbekov2924 4 года назад

    Hi! after installing I have the following error
    ModuleNotFoundError: No module named 'sweetviz'
    what can i do ?

    • @praveenchristopher7776
      @praveenchristopher7776 4 года назад

      same error unable to resolve, any solutions ....

    • @galymzhankenesbekov2924
      @galymzhankenesbekov2924 4 года назад

      @@praveenchristopher7776 so, i finally, resolved it! Basically, what you need to do is to move your folder called sweetviz to the same place as you jupyther file

    • @sourovsahoo7583
      @sourovsahoo7583 4 года назад

      @@galymzhankenesbekov2924 I did what you told but still not working.did u move the whole new environment in which you installed sweetviz to the place we have jupyter files or only the sweetviz file?
      If u moved only the sweetviz file, what is the name of the file?

    • @karthikrams904
      @karthikrams904 4 года назад

      @@sourovsahoo7583 You need to update the PATH in environment variables. I tried it and it worked for me.

    • @sourovsahoo7583
      @sourovsahoo7583 4 года назад

      @@karthikrams904 do u mean i have to change the new created environment path??ping me bro.8763358375.need help

  • @Prakash18996
    @Prakash18996 4 года назад

    Sir u only use python ,I am personally using r for data science

  • @rahulsarkar4794
    @rahulsarkar4794 4 года назад

    Is it better than pandas profiling?

  • @atharv100M
    @atharv100M 4 года назад

    Itz showing invalid syntax i hv used every thing same

  • @sudhanvahabib7582
    @sudhanvahabib7582 4 года назад

    Sadly I didn’t get any response from iNeuron even after filling form prior 6th june :(

    • @krishnaik06
      @krishnaik06  4 года назад

      Hey trying contacting support@ineuron.ai in skype

  • @sagarwaghela1118
    @sagarwaghela1118 4 года назад +2

    Sir From where do you find such things first comment

    • @krishnaik06
      @krishnaik06  4 года назад +7

      From all my subscribers .They tell me all these topics :)

  • @gurdeepsinghbhatia2875
    @gurdeepsinghbhatia2875 4 года назад

    *******SIR EK BAMBOOLIB KAI UPER BE VEDIO BNA DO PLEASE **********
    BAMBOOLIB (VIS LIBRARY)

  • @allieubisse470
    @allieubisse470 4 года назад +1

    Is it not called Data Snooping/ fishing? I mean quickly eyeballing the Test data might lead to Human Bias.
    The algorithm selected might be based on what you saw from the Test data.
    Data Science competitions:
    I see a lot of data Science competitions combining the Train and Test then Do Deep Dive analysis and preprocesing. According to my understanding this is wrong. I personally think that this is done to get more accuracy or overfit the public Leaderboard but i don't think is also applicable to the cooperate world.
    What can we do with our test:
    All we may do is just few checks not going deep to association. Read about Data Leakage.
    Human bias:
    Back to the Human bias, please discuss such. My understanding is to preproces with the Training data and learn some parameters from the training Data then inherit those to test.
    ----------
    I always do train test split before Preprocesing. I make use of sklearn pipeline to avoid the loop of rewriting long code and reduce human error.
    Final word.
    Try Autovis danrothdatascience.github.io/datascience/autoviz.html
    And pyVis

    • @Jam05_
      @Jam05_ 4 года назад +1

      Is it good for large amount of datasets which contains lots of columns and as you said all this visualization tools may lead us to human biasing then why should we use tools like AUTOVIZ

    • @allieubisse470
      @allieubisse470 4 года назад

      To quickly get the overall picture of the training data. Before we can even go deep into tweaking the features.

  • @AshutoshKumar-cj1vj
    @AshutoshKumar-cj1vj 4 года назад +3

    These kinds of features can remove the data scientist role.

  • @sharanbabu2001
    @sharanbabu2001 4 года назад

    Krish checkout %%time in jupyter notebook

  • @syedadil7858
    @syedadil7858 4 года назад

    Wowww. So vizualization tools are worthless now. Lol !