Update stats summary script

[Amy-Xu]

This script has been updated to work with the latest version of Tax-Calculator. Output files get updated as well. Thanks to Martin for pointing out in issue #1562.

[codecov-io]

Codecov Report

Merging #1566 into master will not change coverage.
The diff coverage is n/a.

@@          Coverage Diff           @@
##           master   #1566   +/-   ##
======================================
  Coverage     100%    100%           
======================================
  Files          37      37           
  Lines        2644    2732   +88     
======================================
+ Hits         2644    2732   +88

Impacted Files	Coverage Δ
taxcalc/calculate.py	100% <0%> (ø)	⬆️
taxcalc/taxcalcio.py	100% <0%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 7ef9cb3...ce5a612. Read the comment docs.

[Amy-Xu]

@martinholmer this PR fixed the stats_summary.py in comparison. Could you review?

[martinholmer]

@Amy-Xu said in #1566:

This PR fixed the stats_summary.py in comparison. Could you review?

Thanks. Yes. I'll try to review it no later than tomorrow morning.

[martinholmer]

@Amy-Xu, When I download pull request #1566 to my computer and do the two runs, I get the same results in the variable_stats_summary.csv file, but I don't get the same results in the correlation.csv file. My suspicion is that this is caused by you showing excessive floating-point precision in the correlation.csv file. We already have other evidence that floating point calculations are not exactly the same on Python 2.7 and Python 3.x. And besides, telling users that the correlation coefficient between two variables has a value of, say, -0.0050643581807 is not sensible. Not only is is hard to read a fraction with that many digits, it assigns way too much precision. Why not round the correlation coefficients to three decimal points, so that instead of -0.0050643581807, you show -0.005? Is there any practical difference between those two values?

[martinholmer]

@Amy-Xu, I haven't studied the code in the stats_summary.py script, but I was wondering whether it is possible to have that script produce both output files at the same time. Would it be possible to do that without an excessive amount of work?

[Amy-Xu]

@martinholmer Thanks for the review. I'll round all results to three decimal points.

In terms of generating both files at the same time, the original thought was simply that users may just be interested in one of them so it would be the most convenient for users to produce one a time. Since we also need to update the files, it makes sense to me producing both at the same time. I'll revise the code accordingly.

[martinholmer]

@Amy-Xu said:

Thanks for the review. I'll round all results to three decimal points.

In terms of generating both files at the same time, the original thought was simply that users may just be interested in one of them so it would be the most convenient for users to produce one a time. Since we also need to update the files, it makes sense to me producing both at the same time. I'll revise the code accordingly.

Thanks.

[martinholmer]

@Amy-Xu, Thanks for this update. Look good, so I'll merge it now.
I already updated the RELEASES.md info to describe your enhancement.