This is the
Note
There are (at least) two reasons why following the instructions below wouldn't work:
To build it, you'll first need to run the analysis script, which needs the results from the user study1 to exist in the same directory:
cd analysis
uv run python3 analyze.py
cd ..Then, do the same for the benchmarks (whose data has been included):
cd benchmark
uv run python3 benchmark.py
cd ..And finally, build the master's thesis itself (parallelization here really helps speed up the TikZ compilations):
make -j $(nproc)
We have the following directories:
-
analysis: Contains scripts and data related to the analysis of the user study. -
benchmark: Contains scripts and data related to the technical evaluation. -
content: Contains the actual$\LaTeX$ content of the individual chapters. -
digital: Contains the contents of the digital appendix. -
exposé: Contains the source and PDF of the original Exposé. -
figures: Contains all figures for the thesis, including individual TikZ drawings.
Footnotes
-
These results have not been included here for privacy reasons. Anonymized results are available at
digital/Eval-*.xlsx, but those won't immediately work due to the redacted fields, so you'd need to fill in some dummy data first (or modify the scripts). ↩ ↩2 -
To make the violin plots more readable, I added jittering functionality to tikzviolinplots in this PR. This has been merged by now, but it's not released as part of the package. My thesis still assumes a local version is available, so you need to clone the tikzviolinplots repository and put it in
../tikzviolinplots/relative to the thesis. ↩