alexeiva / intro-data-capstone-musclehub

• 👋 Hi, I’m @alexeiva
• I’m looking to collaborate on type design projects
• Native Cyrillic expert, contact me for consultations
• Cyrillic Script Chair at Granshan Type Design Competition
• How to reach me telegram https://t.me/@alexeiva, a at cyreal.org
• Open-source foundry @cyrealtype http://cyreal.org/

intro-data-capstone-musclehub's Issues

Good job: Contingency Tables

intro-data-capstone-musclehub/Capstone1 - Alexei Vanyashin/musclehub.py

Lines 380 to 381 in bdbfda1

	contingency3 = [[200, 2304],
	[250, 2250]]

Nice job - All of your contingency tables are correct. This demonstrates your mastery of the statistical concepts taught in this course.

Good job: Excellent print formatting

intro-data-capstone-musclehub/Capstone1 - Alexei Vanyashin/musclehub.py

Lines 271 to 280 in bdbfda1

	from scipy.stats import chi2_contingency
	contingency = [[250, 2254],
	[325, 2175]]
	chi2, pval, dof, expected = chi2_contingency(contingency)
	print "\nTest 1. Who picks up an application?"
	print '-' * 60
	print "Chi-2 test p-val: %s" % pval
	print "\nResult is significant? %s" % (pval < 0.05)
	print '-' * 60

This is just a minor comment, but I appreciate the formatting you set up so that your print statements display the chi-square test results neatly and with appropriate labels.

Nice design: Acquisition Funnel Charts

intro-data-capstone-musclehub/Capstone1 - Alexei Vanyashin/musclehub.py

Lines 451 to 467 in bdbfda1

	a3 = (final_member_pivot['Percent Purchase'] * 100).values
	ax3 = plt.subplot(133)
	plt.title('Visitors who purchase a membership', fontsize=14)
	bar3 = plt.bar(b,a3)
	bar3[0].set_color(color_scheme[0])
	bar3[1].set_color(color_scheme[1])
	ax3.set_xticks(b)
	ax3.set_xticklabels(['Fitness Test', 'No Fitness Test'], fontsize=14)
	ticks = range(0, 11, 2)
	ax3.set_yticks( ticks )
	ax3.set_yticklabels([str(s) + '%' for s in ticks], fontsize=12)
	plt.suptitle('Acquisition Funel', fontsize=20)
	plt.savefig('funel.png')
	plt.show()

I think it was a great idea to line up the three different charts you were asked to create side-by-side for easy comparison. It is much easier to analyze the differences between the groups. This is minor, but your green/blue color coding and group text labels both convey the same information of which group each bar belongs to.

Great job: joining query

Great job on crafting a working joining query that pulls together all of the necessary components! The email and gender columns are not actually necessary for the assignment, but it is fine that you have included them.

Summary: Exceeds Expectations

I had a lot of fun going over your project. I was very impressed with the amount of detail and work you put into this project! You put much thought into designing your charts to be easy on the eyes. All of the data visualizations that ended up on your slides are of very high quality. If you are interested in the strategies and thought processes that go into designing data visualizations, I would recommend you check out The Visual Display of Quantitative Information by Edward Tufte.

I won't have too much to say about your code, as you had no errors. All of your python code worked correctly towards accomplishing the various data science tasks. You followed many of the best conventions concerning indentation and variable naming conventions. This helped in making your python code easily readable. Here is more information on python styling conventions if you are interested.

I would suggest you practice adding in your own comments and a README to this github repository to give some context to your code. Comments answer questions like 'What were you thinking when you writing this code?' and 'What is the purpose of this chunk of code? Are there currently any issues with it?" Here is a helpful guide for creating READMEs that give a good introduction to your project.

Your presentation was nicely organized, and you have a logical storytelling flow that keeps readers engaged. I would just suggest taking care with how much information you fit into one slide and making sure not to let the font sizes get too small. On these slides, it became difficult to know what to focus on. You want to make sure that every little thing presented has a purpose and impact since space is limited.Your use of colors and bolding was great in making the slides more visually interesting.

Great job: pie chart

intro-data-capstone-musclehub/Capstone1 - Alexei Vanyashin/musclehub.py

Lines 190 to 202 in bdbfda1

	color_scheme = ['#D4E157', (0.1, 0.2, 0.5, 0.6)]
	plt.pie(ab_counts.first_name.values,
	labels = ab_counts.ab_test_group.values,
	startangle=90,
	colors=color_scheme,
	autopct='%.2f%%'
	)
	plt.axis('equal')
	plt.legend(ab_counts.ab_test_group.values)
	plt.title('Percentage of A and B groups')
	plt.savefig('ab_test_pie_chart.png')
	plt.show()

Your pie chart is excellent and goes above and beyond what was asked for in the assignment. You have a clear title and clear labeling of the A and B groups. I like the colors you chose! You may want to consider that your A and B text labels on either side of the chart and the legend display the same information (it is generally good practice in visualizations to reduce any information that is repeated).

Observation: Styling decisions for plots can make your results look different

This isn't an issue but more of an interesting tidbit to keep in mind when designing data visualizations.
You can manipulate data visualizations to give different impressions.
For example, both plots are created using the same data. Manipulating the range of the y-axis in the following two plots makes the differences between the two groups appear more pronounced in the second plot: