wrote this code quickly, didn't spend any time making it look good, a third is probably copilot

walks.py

import json
import sys
from datetime import datetime, timedelta

def parse(d):
	return datetime.fromisoformat(d)

def proc(d):
	s, e = map(parse, d)
	return e - s

#f = json.load(sys.stdin)

# for all folder in current folder
import os

f = []

for folder in os.listdir():
	if not os.path.isdir(folder):
		continue
	# for each json file
	for file in os.listdir(folder):
		if not file.endswith(".json"):
			continue
		with open(os.path.join(folder, file), encoding="utf-8") as fp:
			try:
				data = json.load(fp)["timelineObjects"]
			except:
				print(file)
				raise
			for obj in data:
				if act := obj.get("activitySegment", None):
					if act.get("activityType") == "WALKING":
						if "distance" in act:
							dist = act["distance"]
						else:
							dist = act["waypointPath"]["distanceMeters"]
						d = act["duration"]
						d = (d["startTimestamp"], d["endTimestamp"])
						speed = (dist / 1000) / (proc(d).total_seconds() / 3600)
						if speed > 15:
							continue
						f.append((dist, d))
			



import numpy as np
import pytz
walk_data = [((dist / 1000) / (proc(d).total_seconds() / 3600), d[0]) 
			 for (dist, d) in f 
			 if dist > 100 and parse(d[0]) >= datetime(2015, 9, 1, tzinfo=pytz.utc) and parse(d[0]) <= datetime(2025, 6, 30, tzinfo=pytz.utc)]
walks = np.array([x[0] for x in walk_data])

med = np.median(walks)
print(med)

print(np.average(walks))


walks = walks[walks > 1]

print(walks)

import matplotlib.pyplot as plt

ax = plt.subplot(111)

ax.hist(walks, bins=45)
ax.grid('on', linestyle='--', linewidth=0.5)
ax.set_xlabel("Walk speed (km/h)")
ax.set_ylabel("Walk count")
# export to high quality png
plt.savefig('walk.png', dpi=300)

# write histo data to json
with open("walk.json", "w") as fp:
	hist, bins = np.histogram(walks, bins=45)
	json.dump({"hist": hist.tolist(), "bins": bins.tolist()}, fp)

# average walking speed per year
from collections import defaultdict

walks_per_year = defaultdict(list) 

for dist, date in f:
	year = (parse(date[0]) - timedelta(days=9*30)).year
	walks_per_year[year].append(dist / 1000)

av_per_year = {year: np.sum(dists) for year, dists in walks_per_year.items() if year >= 2017}

# plot average walking speed per year
plt.clf()

ax = plt.subplot(111)

ax.bar(av_per_year.keys(), av_per_year.values())

with open("yearly.json", "w") as fp:
	json.dump(av_per_year, fp)

ax.grid('on', linestyle='--', linewidth=0.5)

ax.set_xlabel("School year")
ax.set_ylabel("Distance walked (km)")

plt.savefig('walk_per_year.png', dpi=300)

walks_per_month = defaultdict(list)

for dist, date in f:
	if date[0] < "2017-09":
		continue
	adjusted = date[0][:len("yyyy-mm")]
	if dist / 1000 > 10:
		print(date, dist / 1000)
	walks_per_month[adjusted].append(dist / 1000)

av_per_month = {month: np.sum(dists) for month, dists in sorted(walks_per_month.items(), key= lambda x: x[0])}

print(av_per_month)

plt.clf()

ax = plt.subplot(111)

#ax.bar(av_per_month.keys(), av_per_month.values())
# area under line
ax.fill_between(list(av_per_month.keys()), 0, list(av_per_month.values()), color="skyblue", alpha=0.4)

with open("monthly.json", "w") as fp:
	json.dump(av_per_month, fp)

ax.grid('on', linestyle='--', linewidth=0.5)

ax.set_xlabel("Month")
ax.set_ylabel("Distance walked (km)")

plt.savefig('walk_per_month.png', dpi=300)

vals = list(av_per_month.values())
# get 3-month moving average

WIDTH = 3
#WEIGHTS = np.ones(WIDTH) / WIDTH
WEIGHTS = np.array([0.25, 0.5, 0.25])

moving_avg = [np.average(vals[i-(WIDTH//2):i+(WIDTH//2)+1], weights=WEIGHTS) for i in range(WIDTH//2, len(vals) - WIDTH//2)]

plt.clf()

ax = plt.subplot(111)

ax.plot(list(av_per_month.keys())[WIDTH//2:-(WIDTH//2)], moving_avg)

ax.grid('on', linestyle='--', linewidth=0.5)

ax.set_xlabel("Month")
ax.set_ylabel("Distance walked (km)")

plt.savefig('walk_per_month_moving_avg.png', dpi=300)

with open("moving_avg.json", "w") as fp:
	json.dump(dict(zip(list(av_per_month.keys())[1:-1], moving_avg)), fp)

walks_fit.py

import json
import sys
from datetime import datetime, timedelta

def parse(d):
	return datetime.fromisoformat(d)

def proc(d):
	s, e = map(parse, d)
	return e - s

#f = json.load(sys.stdin)

# for all folder in current folder
import os
import xml.etree.ElementTree as ET

f = []

for file in os.listdir("Activités"):
	if not file.endswith(".tcx"):
		continue
	with open(os.path.join("Activités", file), encoding="utf-8") as fp:
		root = tree.getroot()
		for act in root.iter("{http://www.garmin.com/xmlschemas/TrainingCenterDatabase/v2}Activity"):
			# if attribute sport = walking
			if act.attrib["Sport"] == "Walking":
				# get child lap
				lap = act.find("{http://www.garmin.com/xmlschemas/TrainingCenterDatabase/v2}Lap")
				start_time = lap.attrib["StartTime"]
				dist = float(lap.find("{http://www.garmin.com/xmlschemas/TrainingCenterDatabase/v2}DistanceMeters").text)
				time = float(lap.find("{http://www.garmin.com/xmlschemas/TrainingCenterDatabase/v2}TotalTimeSeconds").text)

				d = (parse(start_time).isoformat(), (parse(start_time) + timedelta(seconds=time)).isoformat())
				speed = (dist / 1000) / (time / 3600)
				if speed > 15:
					continue
				f.append((dist, d))

			



import numpy as np
import pytz
walk_data = [((dist / 1000) / (proc(d).total_seconds() / 3600), d[0]) 
			 for (dist, d) in f 
			 if dist > 100 and proc(d).total_seconds() > 300 and parse(d[0]) >= datetime(2015, 9, 1, tzinfo=pytz.utc) and parse(d[0]) <= datetime(2025, 6, 30, tzinfo=pytz.utc)]
walks = np.array([x[0] for x in walk_data])

med = np.median(walks)
print(med)

print(np.average(walks))


walks = walks[walks > 1]

print(walks)

import matplotlib.pyplot as plt

ax = plt.subplot(111)

ax.hist(walks, bins=45)
ax.grid('on', linestyle='--', linewidth=0.5)
ax.set_xlabel("Walk speed (km/h)")
ax.set_ylabel("Walk count")
# export to high quality png
plt.savefig('walk.png', dpi=300)

# write histo data to json
with open("walk.json", "w") as fp:
	hist, bins = np.histogram(walks, bins=45)
	json.dump({"hist": hist.tolist(), "bins": bins.tolist()}, fp)

# average walking speed per year
from collections import defaultdict

walks_per_year = defaultdict(list) 

for dist, date in f:
	year = (parse(date[0]) - timedelta(days=9*30)).year
	walks_per_year[year].append(dist / 1000)

av_per_year = {year: np.sum(dists) for year, dists in walks_per_year.items() if year >= 2017}

# plot average walking speed per year
plt.clf()

ax = plt.subplot(111)

ax.bar(av_per_year.keys(), av_per_year.values())

with open("yearly.json", "w") as fp:
	json.dump(av_per_year, fp)

ax.grid('on', linestyle='--', linewidth=0.5)

ax.set_xlabel("School year")
ax.set_ylabel("Distance walked (km)")

plt.savefig('walk_per_year.png', dpi=300)

walks_per_month = defaultdict(list)

for dist, date in f:
	if date[0] < "2017-09":
		continue
	adjusted = date[0][:len("yyyy-mm")]
	if dist / 1000 > 10:
		print(date, dist / 1000)
	walks_per_month[adjusted].append(dist / 1000)

av_per_month = {month: np.sum(dists) for month, dists in sorted(walks_per_month.items(), key= lambda x: x[0])}

print(av_per_month)

plt.clf()

ax = plt.subplot(111)

#ax.bar(av_per_month.keys(), av_per_month.values())
# area under line
ax.fill_between(list(av_per_month.keys()), 0, list(av_per_month.values()), color="skyblue", alpha=0.4)

with open("monthly.json", "w") as fp:
	json.dump(av_per_month, fp)

ax.grid('on', linestyle='--', linewidth=0.5)

ax.set_xlabel("Month")
ax.set_ylabel("Distance walked (km)")

plt.savefig('walk_per_month.png', dpi=300)

vals = list(av_per_month.values())
# get 3-month moving average

WIDTH = 3
#WEIGHTS = np.ones(WIDTH) / WIDTH
WEIGHTS = np.array([0.25, 0.5, 0.25])

moving_avg = [np.average(vals[i-(WIDTH//2):i+(WIDTH//2)+1], weights=WEIGHTS) for i in range(WIDTH//2, len(vals) - WIDTH//2)]

plt.clf()

ax = plt.subplot(111)

ax.plot(list(av_per_month.keys())[WIDTH//2:-(WIDTH//2)], moving_avg)

ax.grid('on', linestyle='--', linewidth=0.5)

ax.set_xlabel("Month")
ax.set_ylabel("Distance walked (km)")

plt.savefig('walk_per_month_moving_avg.png', dpi=300)

with open("moving_avg.json", "w") as fp:
	json.dump(dict(zip(list(av_per_month.keys())[1:-1], moving_avg)), fp)