Marketing Analytics Project using Machine Learning | Campaign Funnel Optimisation | Project#4

[@skillcate]

Hey guys!! Glad to see such amazing feedback on this ML Project🤗 Need your support in reaching out to more learners by subscribing to my channel 🙂 Also, join me on my Skillcate Discord Server: discord.gg/GyMBfD4ER5 🙂 Let's talk Machine Learning ❤❤

[@nepaltourguide257]

This extraordinary explanation encourage me to do this comment....really 👏 outstanding video...

[@skillcate]

Thank you so much 😀

[@AbhishekChandraShukla]

This is a serious marketing mix modeling video.

[@skillcate]

Glad you think so! Please share in your network if you think they will find this useful too!

[@AbhishekSingh-xz3nk]

Superb video ❤ Detailed and interesting to watch

[@skillcate]

Glad that you find it useful!

[@AbhishekChandraShukla]

Bro this is actually helpful! Thank you, brother. Really appreciate it.

[@ArshuVaida]

Great one, well explained.

[@method341]

I haven't watched this fully but thanks for uploading this. It looks great!

[@skillcate]

Glad that you liked. Do subscribe to your channel for more such amazing content.

[@AbhishekChandraShukla]

Commenting again, bro! Good job.

[@skillcate]

You're the best!

[@schoolsami]

Thank you for everything. I'm a student, I'm learning about machine learning. I have finished my project. Currently, I want to deploy a Machine learning project to the Website. But I don't know how to do it, do you have any video or can you help me.

[@skillcate]

Hey, hope you enjoyed finishing your ML project 🙂 There are couple of videos I have done on deployment that would help you: 1. Complete guide on deploying a Python Project using Flask: kzbin.info/www/bejne/n6eYaamZibaUqqM 2. Deploy a Time Series Flight Fare Prediction Project (similar to Marketing Strategy) with Flask App: kzbin.info/www/bejne/nKKVpYuBiNKrbNU Share your feedback on how it went..👍

[@schoolsami]

@@skillcate I have a question, why when using Logistic, the amount of "10% data" goes from 22K to 4K data. And the amount of "data 90%" is only 1000 data. What is the "ClusterGroup" data A,B,C,D? What is the meaning of A, B, C, D? and "TVReg"? Sorry if my question feels stupid, because I'm a student I don't have much experience.

[@tifeogunjimi5904]

Awesome as ever. Well done. Quick question though, How will the business then know specifically which customers are in each Decile? since we dropped the customer IDs and the customer IDs are not included in the model and prediction outputs? Thank you for all your help.

[@skillcate]

Great point Tife!! I have updated the prediction code file: b2_Predictor_Marketing.ipynb, to include Customer ID column in final output file. Check it here: drive.google.com/drive/folders/1CD1XBknEICitmbfEi0XR2cNox3oyQIMJ?usp=sharing

[@swetasharma8467]

Thank you for such a great video. Extremely Insightful!

[@skillcate]

Glad it was helpful!

[@wesolvedigital_]

what was that at 23:15 ? Which commands did you use?

[@shirinkhan3004]

this was actually very detailed and organised thank you so much

[@skillcate]

Glad you liked! Happy learning!

[@mubasheryousuf1771]

Doing such a great job Thanks for all Projects

[@skillcate]

Thanks Mubasher for the feedback! Hope you found them helpful!

[@jpaokx]

Great video....do you have any video around channel attribution?

[@skillcate]

Hey buddy!! Thanks for your reply. However, I would need slightly more info around this one to help you further. :)

[@JiunyuHsieh]

Very useful and organized video! Thank you so much!! May I ask a question? Why did you use mode to replace most of the null values but choose mean for the 'LoyalTime' Field

[@skillcate]

Dear Jiunyu, Thank you for your kind words and for watching the video! I'm glad you found it useful and organized. And it's a great question you asked. Well, in the machine learning project, we used the mode to replace most of the null values because the mode represents the most frequently occurring value in a dataset. It is commonly used for categorical or discrete variables where the concept of "most common" makes sense. However, for the 'LoyalTime' field, we chose to use the mean to replace the null values. The reason behind this decision is that 'LoyalTime' is a continuous numerical variable that represents the amount of time a user has been loyal. Using the mean allows us to approximate the average loyalty time of the users with missing values, providing a more representative estimate. Hope it's clear now :)

[@JiunyuHsieh]

@@skillcate That is helpful! Thank you for your reply😀

[@mohamedanuarbinibrahim906]

Thank you for providing great tutorials. Can I ask how did you get the profit of 214 and 196 mn inr in your final report when the model output analysis in excel shows 4.2mn and 3.8mn?

[@skillcate]

Dear Mohamed, hope you are doing well!! Good job in pointing out this details, buddy :-) Model Output Analysis (based on Decile Methodology) is done on ~20% test data (which is ~4500 observations). So, the profit numbers there are 4.2 and 3.8Mn. As we have ~450 participants here per decile, so total participants are low, so the final numbers are also low. However, in our final report, we talk about our problem statement where we need to build a strategy on how do we reach to 225000 loyalty program participants (the 90%). I have updated the Final Report with the formulas now. Calculations should be clearer now. Let me know if you have any further doubts. Thanks for the valuable feedback :-)

[@akshanshbaliyan]

@@skillcate Hi. There is still some confusion on how you transformed these number for 90% dataset. 1. How come 90% data has 1000 observations only? 2. Your cumm good % is not matching up with your Model Output Analysis. Thanks in advance.

[@varunsingh4369]

Thank you for the video but can you provide the link for dataset from where you have taken this... I have to mention it in my assignment.

[@skillcate]

This dataset is prepared by us through some reference to IBM SPSS Academic Training Module. You may read more on SPSS datasets on this link: www.ibm.com/docs/en/spss-statistics/24.0.0?topic=system-sample-files

[@Vlapstone]

my friend, where did you take this 4.1 output dataset from? how do you make this table?

[@skillcate]

Hi Liciano, the Output Analysis file is prepared from the CSV File we wrote towards the end of our code. Here's the Google Drive Link that has all the files: drive.google.com/drive/folders/1CD1XBknEICitmbfEi0XR2cNox3oyQIMJ

[@Vlapstone]

@@skillcate Thanks a lot!

[@ParasBansalPaaru]

Amazing. Looking forward to more such case studies.

[@skillcate]

Thanks Paras for the feedback. We have recently launched couple of new project, one on: Age, Gender & Emotion Detection (kzbin.info/www/bejne/q6DZoGSpZseUqLc) and second on: Credit Scoring (kzbin.info/www/bejne/bpvdp62IpJh8oJI&t)..😊

[@SandeepSingh-ul1pl]

Great video. Can you list the source of the dataset ?

[@skillcate]

This dataset is prepared by us through some reference to IBM SPSS Academic Training Module. You may read more on SPSS datasets on this link: www.ibm.com/docs/en/spss-statistics/24.0.0?topic=system-sample-files

[@hassan.razzaq]

#Testing 31:02