12. Project - 8 (Case Study - 8) | Data Analysis With Python Pandas | Kaggle Dataset

[@shakilahammed1887]

Thanks a lot

[@abhishekanand5898]

Awesome sirji

[@DataThinkers]

Thanks for your comment. Glad you like this video.keep watching.

[@manojhaygale810]

good work and very understanding thanks

[@DataThinkers]

Thank you for your comment. Glad you like this video. you can watch other pandas projects from the following playlist. Link : kzbin.info/aero/PL_1pt6K-CLoDMEbYy2PcZuITWEjqMfyoA Thanks.

[@philtoa334]

Nice.

[@DataThinkers]

Thank for your comment. Glad you like this video.

[@sobatbale4569]

Love it

[@DataThinkers]

Thank you for your comment. Glad you like this video.

[@utkarshkunwar4319]

Hello Priyang, thanks for all these tutorials! They indeed help a lot for practice. Although I've some differences with you in this particular video, right from question number 14 till 17 won't we have some different graphs from what you have since Grade 'A++' or Grade '5' have only 10 appearances hence they won't be the longest bars in any of the graph that we plotted. 4 I would really appreciate your answer to this, thank you and take care!

[@nikhil_somani]

Hi, apperance of one parameter cannot define other parameter as the corresponding value will make the actual sense. Like even having less entries of Grade 5 if it is on top level it can be checking out the other parameter value given. we can have 10 bags of 50 gm each making a total of 500 gm only while can have 1 bag of 50 kg, making a total of 50 kg price 50 kg have more value and if have same thing in all bag. I hope it helps. I might be wrong some way. Please correct me if so. thank you.

[@frolilapume]

First, thanks for the amazing content! But I think the logic behind the percentage of missing values in exercise 7 is wrong. When you use len(data) you get the number of rows in the dataset which is 5000. But to compare the number of missing values with all values of the dataset, you'd have to get the total number entries (values) in the dataset which would be (data.count().sum())+(data.isnull().sum().sum()) First I get the number of entries data.count().sum()) (not counting the missing values, because .count() is not considering NaNs) and then I add the missing values count data.isnull().sum().sum(). This results in 30000 entries, which makes sense since we have 5000 rows and 6 columns in the dataset. And from here we can make the percentage, that would look more like this: (data.isnull().sum().sum()) / (data.count().sum())+(data.isnull().sum().sum()) * 100. The result is 1.31 percent, which makes more sense to me since we have 30000 entries in total, and 393 NaNs. Please correct me if I'm wrong. Cheers.

[@dpegn]

Just curious which machine learning model will be used in this case? As we have given numbers 1- 5 to grades ?

[@deep_0212]

why do we use regex=True in the replace function?

[@Practicaltheoretical]

Sir please make more videos

[@DataThinkers]

Sure

[@ridoychandraray2413]

Bro we want more videos by pandas

[@haniabdulrehman7774]

Hello, Thanx for ur tutorials. But sometimes i am facing problem in pandas after i convert a column(like Grade column)to dictionary values i assign but instead of those values 'Nan' is being shown.. It is like jupiter is crashing & i'm not able to do further problems further. Can you plz help me in this regard?

[@DataThinkers]

Check this code:github.com/DataThinkers/Data-Analytics-Projects-Code/blob/main/Project%20-%208%20KZbin%20Channel%20Dataset.ipynb

[@m.mukeshmugan7761]

from question 14 to 17 you ploted graph but i was looking to answer that question by numerical values .when should i use graph and when numerical value as answer

[@DataThinkers]

You should use numerical values as an answer when you need to provide precise and specific information or perform calculations. Use graphs (e.g., plots, charts) when you want to visualize trends, patterns, or relationships in the data to make it easier for others to interpret and understand. Graphs are especially useful for conveying complex information and aiding in decision-making. Choose the appropriate approach based on the context and the goal of your analysis.

[@heshamsbr]

I tried to use this command for the dataset but it turned out no rows "df[df['Subscribers']=='--']"!!!!Can you advise the reason

[@ishani5171]

because -- is a non-word character and inside " " a string is supplied

[@apensquared]

Hello Priyang, thank you for making these videos! Just one suggestion - if you can stop repeating "so now", and "as you can see" soo many times, in my opinion, your videos will be easier to listen to.

[@DataThinkers]

Actually, I'm habituated to speaking like this, but I'll certainly give it a try. Thank you.

[@yasmeenkarachiwala9612]

Hello Sir, Why does the groupby yield a different output compared to barplot? df_7.groupby('Grade')['Video Uploads'].sum() Grade 1 8536620 2 4124010 3 5121740 4 678412 5 374507