I’m happy to share with you my another guided project which took me almost the whole month to finish I worked a lot on it and tried to incorporate all the knowledge that I obtained writing my articles. Namely, I used missing value visualizations, stem plots, pretty-printed tables. I paid a lot of attention to the dataviz part making each plot as readable and informative as possible. Also, I optimized my code by introducing many case-specific functions and (for the first time) added descriptive docstrings to each. Apart from considering best markets by country, I analyzed also the situation by age categories, but only for the USA, since it’s a predominant market anyway.
An interesting thing that I noticed is a huge amount of people willing to learn to program free of charge. Well, probably not so surprizing: the dataset in use is the one of a freeCodeCamp survey, and this website offers free learning. In my project, I experimented with excluding such people from the analysis, but the overall proportions (the number of learners by country and by age category) remained almost the same.
Another extremity is represented by people who learn to program for scaringly high prices, up to $80,000/month! I beleive that there is always a limit to the quality/price relation, especially considering that a lot of people managed to learn new skills in whatever sphere almost free of charge. Well, as for learning Data Science, I’d definitely advise them DQ, as a way to save their money and to have a rapid and high-quality progress in their learning!
I’ll be very grateful to you for reviewing my project and giving your feedback. What can be improved in terms of the code or storytelling? Are their some discrepancies, mismatches, errors, or just typos? What can be skipped or added? Whatever thing you’ll notice and let me know would very helpful to me!
Thanks a lot in advance!
Finding the Best Markets to Advertise In.ipynb (1.3 MB)
Click here to view the jupyter notebook file in a new tab
Am very glad to interact with another project of yours on Finding the Best markets to Advertise In.
Have took couples of hours going through your project and honestly have gained a lot. On realizing that it took you almost a whole month to finish the project, I was so curios about this, and decided to fully engage my understanding in every step which in return paid off . I ended up noting down some of the code lines that I interacted with for the first time and were so interesting .The information given in the introduction, are so awesome, the exposure given on E-learning , the aim are just excellent. I love how you have summarized the results.
Information background of the dataset has been well worked on, the links provided, general information about freeCodeCamp is just wonderful.
Being that haven’t tackle this project, the doctrines you included in the functions help me to understand most of the functions that appeared to be complex. The explanations you gave on which columns to be retained are awesome.
While going through the analysis I noticed some interesting plots like the stem plots, well printed tables and so on, all these motivated me to get into details of what the output entails ,facts were never hidden. I love the way you dealt with the outliers, the explanations given are so recommendable, the conclusion is so informative and nicely presented.Just want to thank you mate for sharing your project, which from the beginning has been helpful to most of the learners.
kindly check on the auto-generated link( for viewing purposes of the project) , it’s returning an error, I think reuploading the project will solve the problem. And is `missigno’ supported by Anaconda? have tried working with it locally in my machines but ‘module not found error’ was returned, or is there additional libraries I need to have in my machine?.
Thanks a lot for your detailed feedback and kind words, I’m very excited that my hard work is appreciated and helpful for other students! I’m always trying to do everything as flawless as possible (well, in my case, I currently have also a lot of spare time for it ), so I’m happy that at least the result is good.
As for missingno, yes, you have to install first this library (
pip install missingno), it’s not included in Anaconda. Then, already in Anaconda, you need to import it (
import missingno as msno). It’s a wonderful tool to visualize missing values! It doesn’t actually have many graph types (practically, only 4) and the customization is somewhat limited. But anyway, all its graph types, settings, and methods are more than exhaustive for creating insightful visualizations.
Stem plots are also a cool (and highly customizable, if necessary) alternative to bar plots, especially when there are many bars, or if they are of similar length. In these 2 cases, stem plots look much more elegant and digestible than bar plots, and they are definitely better in terms of data-ink ratio. And for using them, you don’t have to install or import anything extra.
Oh, these faulty project links! I also noticed that something is going wrong with them. I tried to fix it manually (both the link to my project and those to the other people’s projects which I was going to review yesterday), but it was impossible, and always threw the same error! Then, I tried to create a ticket, but also the ticketing system currently doesn’t work. I think that there is some technical system updating ongoning on DQ, so I’ll wait probably till tomorrow and if nothing is resolved, I’ll contact DQ staff about it.
Thank you very much again for reviewing my project!
Thank you very much @Elena_Kosourova for further clarification on stem plots, missigno and even the faulty project links. Happy codding
Brayan, I have just checked, the links on the projects work again, both on mine and on other people’s projects! And the ticketing system (I checked also it) still doesn’t work for now, hopefully will be fixed soon as well.