scripts and generated / edited data for my Bachelor's thesis at HSLU spring semester 2024, concerning a possible directional positioning error in smartphone GPS

new data can be generated by using the Android app that was also part of this bachelor's thesis: Measuralyze

• data_profiling.py: used in data quality assessment for sanity check to make sure data makes sense as well as find errors with some general analysis

• calculate_distance.py: calculates distance using haversine formula, used to generate the distances between actual location and gps location that is needed for analysis

• boxplot.py: used to generate box plots of entire data

• correlation_numerical.py: calculates the correlation for numerical attributes and target attributes latDifference, lonDifference and distance using pearson correlation

• correlation_boolean_categorical.py: creates boolean attributes for categorical attributes, in the style of vector space model, then calculates the correlation for boolean attributes and target attributes latDifference, lonDifference and distance using point-biserial correlation

• polar_plots.py: plots the direction of each error for all locations (and stages, they are shown in chapter 6.3.3).

• plot_distances_for_series.py: splits up the data into series of 20 measurements and plots graphs for series

• first_measurement_vs_median_measurement.py: calculates the difference between distance of first measurement in series to actual location and median measurement to actual location, then plots them

• polar_plots_video.py: creates a couple more polar plots with different styling (and different titles) that were used in the video pitch