After almost 3 weeks of work (2 hours a day, on average) I have finished the project (still struggling on how to take additional steps besides the DQ instructions and finish the project faster than in 2 weeks ). Many times IĀ“ve got the feeling that I dig into too many unnecessary details while working on the guided projects, which can be somewhat demotivating sometimes. So, besides the ways of improving the project IĀ“m looking for sincere answer on following question: do the additional steps taken reveal some interesting insights or is it digging in for digging in?
@ksenia.kustanovich Wow, I have to say your effort has definitely paid off. Your project looks absolutely amazing, thorough, and professional looking. I especially LOVE the plots you created.
I know what you are talking about with ādigging inā too much. I have felt that way with almost all of the guided projects. From my own experience, I learn the most when I dig deeper and go off the rail from the guides. Cause Iām motivated enough to find my own answers. Since we are learning right now, I donāt think digging deeper is going to cause any harm, most of the time, quite the contrary.
But then again, we do face the question of balancing how much deeper do you dig, and how much time do you spend on a project. Thatās why I think itās important to have a clear goal for each project. It may not be clear at first, and it may change as you explore the dataset more, but itās helpful to keep the goals in mind. What I find helpful is to list the questions I want answers to from a project at the beginning, and you can always modify them as you go.
Again, great job on the project, I really admire the diligence and the time and thoughts you pour into this project.
Judging by last couple of projects, I realize that I do enjoy working on the visualisation part of the projects, starting from choosing the mos adequate type of plot and then transforming the data to fit it to the plot and the aesthetics part as well.
The truth is that IĀ“m quite satisfied with almost all of my projects, especially with the things I learnt in the process beyond the DQ curriculum. But then I realize that my progress on the DA path is quite slow and feel somewhat upset.
You are so right about finding the balance. Maybe in addition to the questions IĀ“d like to answer I should also specify the deadline for the project and stick to it while deciding what new ideas that come up in the proccess I want to pursue.
Thanks one more time. IĀ“m really glad that you liked the project and find it worth 3 week work on it
Hello @ksenia.kustanovich! Your project is just amazing, I would say itās a great candidate for the best shared project this week.
Itās very good that you decided to get beyound the DataQuest curriculum and work your own way. Itās very slow at the beginning (as it should be) but once you get used to do things well it will be much easier and quicker to create profesinally looking projects like this one.
I especially liked your work on data viz. Good choice of plots to use and colors make your plots very clear with no need to go to deep explanations!
However I would give your some tips on improving further your plotās style:
Align the titles with the x axis labels
Make sure that you label x and y labels even when it seems that they are self-explainable (it really helps the reader to understand every sing part of your plot, you just do not leave space to misinterpretation)
Thank you, Artur! Both you and @veratsien have cheered me up.
Thank you for your tips as well. Iāve got a question though:
[quote=āartur.sannikov96, post:4, topic:546814ā]
Align the titles with the x axis labels
[/quote] - could you please explain what you mean may be with some example?
[quote=āartur.sannikov96, post:4, topic:546814ā]
Make sure that you label x and y labels even when it seems that they are self-explainable (it really helps the reader to understand every sing part of your plot, you just do not leave space to misinterpretation)
[/quote] - I thought it was better to leave out some labels in order not to distract from the plot itself in cases where their meaning was obvious. So, Iāll remember this tip next time I work on a plotās aesthetics
Once again, thank you for your comments! Iām glad that you liked the project.
As for the labels: in some cases itās more of a personal choice, if you see that your data is self-explainable, do not use them. But for example on the KDE plot at the end, at least the numbers on x axis are not very clear for meā¦
An excellent project, the love for the details is astonishing, off course all this take a lot of time an iteration, and change this and that, congratulations.
Regarding if is worth the effort to explore further, I think is ok if the results are showing something that you didnāt expect to find, and is solid enough to take the chance and explore further. Nothing more to say, thank you so much for sharing!
Your project looks just perfect! Very profound analysis (I definitely wouldnāt call it digging for digging), interesting insights, clear and exhaustive storytelling, well-commented and clean code and, especially, awesome visualizations. My favourite visualizations in your project are the catplot and the bubble-maps, it was a great idea to combine 2 factors on one map. As for the analytical part, I especially liked your idea to investigate racially-balanced schools and their academic results, and also very interesting insights about the cases where students or teachers gave the highest scores in the survey.
Here are some suggestions from my side:
The heatmap correlations: it would be better to swap the colors (the red for positive correlations, the blue - for negative).
The fontsize on the plots, especially on my favourite catplot , should be increased, otherwise some notations are difficult to read.
Itās better to use a uniform quote marks style for the string data throughout the project (only single or only double).
The code cell [5]:
def padded_csd(val):
val=str(val)
return val if len(val)>=2 else val.zfill(2)
This zfill method doesnāt actually require checking the current length of the value. It adds zeros where necessary and leaves the other values as they are. So this function can be written simply:
Thank you very much for such a detailed feedback!
ItĀ“ll motivate me not to lower the level for my further projects.
Glad that you noticed the racially-balanced schools investigation, it `s one of the parts of the project IĀ“m proud of.
As for your suggestions:
I see your allegation to the temperature colors. It might take place.
The truth is that I failed to change the figure size and the font sizes for this plot IĀ“ll have to investigate it further before adding it to my profile.
I absolutely agree with you. I copied a couple of code blocks from the solution notebook (to speed up a little bit and forgot to change the quote marks style.
thank you for pointing it out. I didnĀ“t know about it. Actually I did this part according to the DQ instructions in one of the missions of the same course.
One more time, great job reviewing the projects. It totally deserves the recognition as the Com.Champ. you get second week in a row:)
Thank you, Ksenia! I also learn a lot when reviewing high level and well-structured projects like yours. Looking forward to seeing your next works. Iāve become your fan now!
). Many times IĀ“ve got the feeling that I dig into too many unnecessary details while working on the guided projects, which can be somewhat demotivating sometimes. So, besides the ways of improving the project IĀ“m looking for sincere answer on following question: do the additional steps taken reveal some interesting insights or is it digging in for digging in?
Very profound analysis (I definitely wouldnāt call it digging for digging), interesting insights, clear and exhaustive storytelling, well-commented and clean code and, especially, awesome visualizations. My favourite visualizations in your project are the catplot and the bubble-maps, it was a great idea to combine 2 factors on one map. As for the analytical part, I especially liked your idea to investigate racially-balanced schools and their academic results, and also very interesting insights about the cases where students or teachers gave the highest scores in the survey.
, should be increased, otherwise some notations are difficult to read.
IĀ“ll have to investigate it further before adding it to my profile.