Looking for feedback: NYC high school data

Hi everyone,

it would be great to get your feedback on my guided project. :slight_smile:
Academic success in high schools: What drives differences in SAT-scores across NYC?

In addition to the scope of step 2 of the data analyst learning path, I have explored the following topics:

  • interactive charts with pyplot and cufflinks
  • interactive maps with folium
  • writing my own function to expedite data exploration (‘describe_plus’) - see my question below
  • accessing data through APIs using sodapy
  • reverse geocoding with geopy

I am especially looking for your thoughts on how you compare a subset (in this case for example schools in Brooklyn) to the main dataset. I found it cumbersome to code several plots - especially given the vast number of features of this dataset. What is your general approach?

Many thanks

Link to final mission screen


Hey @tim1albers

Great project, love the different visualization techniques you have used.
It is a great read.

It is cumbersome to code several plots, but more the plots, more intriguing the project becomes. Generally, if there are too many plots, one could always use more sophisticated tools that minimize coding inputs.

I haven’t done this project yet, hence I am unable to give more technical feedback sorry.

Raj Tulluri


Hi Raj,
many thanks for the kind words. I also really enjoyed your projects on predicting car prices - very thorough!


1 Like

Hey @tim1albers
Thanks for taking out the time to read it.

hey @tim1albers

I haven’t worked on plotly at all, so now your project will be start and reference point for me personally. And even I haven’t worked on basemap. I tried it’s replacement Cartopy but I really couldn’t work it out so this project is overdue for me.

I will break your project into three parts:

Narrative - this is definitely a great read, the sections, sub-sections and descriptions are detailed and explain the process really well. Also the detailed comments for codes where you have added your own ideas or deviated from the given instructions. :ok_hand:

Plots - well no comments. I would have done this in tableau but now I will learn it from here and the sources you have attached as to how to do it in Plotly! :+1:

Technical - Well here I will say your project is too rich and it requires it’s time to be read through otherwise it won’t do any justice to the project. so the project is now bookmarked for my own learning. :+1:
One thing I do wish to highlight, you could break the code cells further into smaller pieces. you may think this would make the project lengthy but then the sub-sectioning would be about what exactly the below code does. and it would add more readability to the code. just a thought.

Of course the “describe_plus” takes the cake! How do you guys have these brilliant ideas to highlight a dataframe like this? :scream: :thinking:

Coming to your query about multiple plots (@Raj has already highlighted certain aspect), I wish to say this -

It depends on what exactly the objective of the project is. Say if we are only interested to understand the SAT Score vs. Certain Demographic (hispanic/ native/ people of color :frowning: this still sounds derogatory) then we can filter out the data to only consider columns/ rows related to these 3 dimensions and create a project. Just like how DQ sometimes gives us a cleaned data set to work with.

But if the project’s objective is to look at multiple aspects then you can’t just jump from objective to the conclusion. You will have to create a start, a center, an end, and a summary for all 3 together. You may, of course, choose that if a similar analysis is given in the GP, you can infuse your own ideas (like you did!) or you can take one aspect of analysis and have deep dive in it (provided you have sufficient data to do so). Also, how would you clearly show what all you have learned to become a data-scientist or data-analyst or your preferred role.

Imagine me coming to your project and post this instead of writing all of the above:

cool project :stuck_out_tongue_closed_eyes:

Of course, that would keep you awake - I can imagine you dozing off in-between while reading my write-up - but it won’t bring any significant value to your project nor will it help us both in our learning!

Thank you for sharing your project with DQ Community! and looking forward to learn more from your projects. :slight_smile:


Hi, @tim1albers. I really appreciated your project. Great visualizations, I love them.

There is so many to learn from your project. I will definitely bookmark it as a learning resource! :blush:


Hi @Rucha,
thank you so much for taking the time to inspect the project and crafting your detailed feedback - much appreciated!
I will definitely keep your feedback on breaking code into more cells in mind on future projects. I will also reflect critically on the comprehensibility of the results I am presenting - you are making a very valid point in terms of effective storytelling!

Many thanks again and best regards,

1 Like

Hi @artur.sannikov96,
many thanks :slight_smile: I tried to address your feedback from the Exit Survey Project, in terms of sufficiently commenting and discussing results.


1 Like

Hi @Raj,
thanks again for your feedback. I looked into tools to minimize coding input and landed on nbextension widget which I have used in the Star Wars survey project.

Thanks for the nudge :slight_smile:

1 Like

Hey @tim1albers

That’s wonderful. Cheers!!

1 Like

hi @tim1albers

forgot to add this earlier, I see a gap after code cell 8. this is on 50% zoom of edge browser.

Edge is my default browser for DQ. so haven’t checked it on others. what exactly is here a chart/plot/table that NBviewer doesn’t render or some other issue?

Hi @Rucha,

on my machine running Safari it look like this:

Best Tim

1 Like

I love the way you have presented the work and it’s certainly a good reference for interactive plotting. Thanks for putting so much effort into this work. It inspires. Great work Tim!


Thank you very much @enext7771 :slight_smile:

Awesome work! There’s definitely a lot of content to learn from in this project. Also a wonderful reference point and a source of inspiration for all of us. Thanks for sharing! :smiley:


Thank you @menon.aj! :slight_smile:

1 Like

Hi @tim1albers,

This is just amazing. I’m learning a lot just by going through your very organised project.
I hope you don’t mind me adopting some of the functions that you’ve written down.
Your use of plotly and folium also steered me to begin exploring those 2 awesome packages.

Thank you!


Great! It looks amazing and I definitely will try to learn some tricks from there! Thank you!!

1 Like

I commend you for the efforts put into this, the visualization is top notch.
I tried using the folium like you did but couldn’t get a hang on it.
Was not telling me module has no Map.
Could you please be of help as regards how you went about it?

1 Like

Thank you for going above and beyond what we’ve learned up to this point in the course series. I’ve learned a good deal from you and will now plagiarize some of your code, engulfing your methods and morphing them into my own.


1 Like