feat: initial DVF data pipeline with 6-level price aggregation

.gitignore

@@ -0,0 +1,48 @@
+# ---- Data (raw and intermediate - too large to commit) ----
+data/raw/
+data/processed/
+# NOTE: data/aggregated/ is NOT ignored - those JSON files are needed by the app
+
+# ---- Python ----
+__pycache__/
+*.py[cod]
+*.egg-info/
+dist/
+build/
+wheels/
+.eggs/
+*.egg
+.venv/
+venv/
+
+# ---- Jupyter ----
+.ipynb_checkpoints/
+
+# ---- Environment / secrets ----
+.env
+.env.*
+
+# ---- Claude Code / MCP / IDE ----
+.claude/
+.mcp/
+.vscode/
+.idea/
+*.code-workspace
+
+# ---- Working notes (not part of the deliverable) ----
+updates/
+
+# ---- Reference repos (downloaded for research, not part of project) ----
+explore.data.gouv.fr/
+data-gouv-skill/
+datagouv-mcp/
+stats-explorer-datagouv/
+
+# ---- OS ----
+.DS_Store
+Thumbs.db
+
+# ---- Misc ----
+*.log
+uv.lock
+.python-version

data/aggregated/prices_country.json

@@ -0,0 +1 @@
+{"FR": {"tous": {"median": 2307.7, "mean": 2953.2, "q1": 1361.4, "q3": 3690.4, "volume": 4735156, "confidence": 0.6}, "appartement": {"median": 3106.7, "mean": 3878.9, "q1": 1985.3, "q3": 4742.4, "volume": 1987882, "confidence": 0.645}, "maison": {"median": 1858.8, "mean": 2266.8, "q1": 1112.6, "q3": 2900.8, "volume": 2549352, "confidence": 0.615}, "local": {"median": 1597.9, "mean": 2497.1, "q1": 827.0, "q3": 3000.0, "volume": 197922, "confidence": 0.6}}}

data/aggregated/prices_department.json

@@ -0,0 +1 @@
+{"01": {"tous": {"median": 2250.0, "mean": 2545.7, "q1": 1500.0, "q3": 3278.7, "volume": 44534, "confidence": 0.684}, "appartement": {"median": 2380.0, "mean": 2771.5, "q1": 1630.9, "q3": 3823.5, "volume": 15712, "confidence": 0.631}, "maison": {"median": 2250.0, "mean": 2488.5, "q1": 1500.0, "q3": 3125.0, "volume": 26914, "confidence": 0.711}, "local": {"median": 1098.4, "mean": 1493.8, "q1": 645.6, "q3": 1839.5, "volume": 1908, "confidence": 0.6}}, "02": {"tous": {"median": 1105.3, "mean": 1215.1, "q1": 749.5, "q3": 1545.5, "volume": 33145, "confidence": 0.712}, "appartement": {"median": 1197.9, "mean": 1298.5, "q1": 921.9, "q3": 1588.8, "volume": 3910, "confidence": 0.777}, "maison": {"median": 1092.1, "mean": 1186.5, "q1": 737.8, "q3": 1535.7, "volume": 28134, "confidence": 0.708}, "local": {"median": 877.2, "mean": 1648.9, "q1": 459.8, "q3": 1777.8, "volume": 1101, "confidence": 0.6}}, "03": {"tous": {"median": 971.7, "mean": 1111.3, "q1": 613.2, "q3": 1437.5, "volume": 27890, "confidence": 0.661}, "appartement": {"median": 1121.5, "mean": 1249.1, "q1": 724.6, "q3": 1578.9, "volume": 7543, "confidence": 0.695}, "maison": {"median": 926.1, "mean": 1058.1, "q1": 588.2, "q3": 1383.5, "volume": 19418, "confidence": 0.657}, "local": {"median": 735.3, "mean": 1105.8, "q1": 416.7, "q3": 1270.5, "volume": 929, "confidence": 0.6}}, "04": {"tous": {"median": 1987.2, "mean": 2106.6, "q1": 1363.6, "q3": 2681.7, "volume": 17062, "confidence": 0.735}, "appartement": {"median": 1890.6, "mean": 1987.9, "q1": 1367.1, "q3": 2493.3, "volume": 7888, "confidence": 0.762}, "maison": {"median": 2170.0, "mean": 2270.8, "q1": 1443.8, "q3": 2891.6, "volume": 8453, "confidence": 0.733}, "local": {"median": 1118.9, "mean": 1480.4, "q1": 687.5, "q3": 1880.0, "volume": 721, "confidence": 0.6}}, "05": {"tous": {"median": 2418.4, "mean": 2596.6, "q1": 1750.0, "q3": 3222.2, "volume": 19480, "confidence": 0.756}, "appartement": {"median": 2473.2, "mean": 2661.5, "q1": 1875.0, "q3": 3254.0, "volume": 12155, "confidence": 0.777}, "maison": {"median": 2433.3, "mean": 2618.8, "q1": 1536.0, "q3": 3333.3, "volume": 5287, "confidence": 0.705}, "local": {"median": 2049.6, "mean": 2151.6, "q1": 1324.8, "q3": 2677.1, "volume": 2038, "confidence": 0.736}}, "06": {"tous": {"median": 4347.8, "mean": 4709.7, "q1": 3220.9, "q3": 5659.1, "volume": 138875, "confidence": 0.776}, "appartement": {"median": 4444.4, "mean": 4801.3, "q1": 3423.9, "q3": 5684.2, "volume": 110545, "confidence": 0.797}, "maison": {"median": 4000.0, "mean": 4622.5, "q1": 2520.2, "q3": 5769.2, "volume": 21801, "confidence": 0.675}, "local": {"median": 2900.0, "mean": 3449.3, "q1": 1829.3, "q3": 4400.0, "volume": 6529, "confidence": 0.645}}, "07": {"tous": {"median": 1634.1, "mean": 1786.4, "q1": 1067.6, "q3": 2338.2, "volume": 23768, "confidence": 0.689}, "appartement": {"median": 1384.6, "mean": 1510.9, "q1": 959.0, "q3": 1916.3, "volume": 6004, "confidence": 0.723}, "maison": {"median": 1804.0, "mean": 1921.8, "q1": 1178.6, "q3": 2481.8, "volume": 16726, "confidence": 0.711}, "local": {"median": 918.6, "mean": 1198.0, "q1": 533.4, "q3": 1500.0, "volume": 1038, "confidence": 0.6}}, "08": {"tous": {"median": 1045.5, "mean": 1154.5, "q1": 684.4, "q3": 1506.8, "volume": 16526, "confidence": 0.685}, "appartement": {"median": 964.3, "mean": 1022.0, "q1": 684.8, "q3": 1288.7, "volume": 3070, "confidence": 0.75}, "maison": {"median": 1089.1, "mean": 1185.5, "q1": 702.5, "q3": 1581.2, "volume": 12847, "confidence": 0.677}, "local": {"median": 692.3, "mean": 1169.0, "q1": 432.4, "q3": 1155.2, "volume": 609, "confidence": 0.6}}, "09": {"tous": {"median": 1277.1, "mean": 1384.5, "q1": 833.3, "q3": 1785.7, "volume": 14126, "confidence": 0.702}, "appartement": {"median": 1296.7, "mean": 1383.2, "q1": 893.8, "q3": 1704.5, "volume": 2343, "confidence": 0.75}, "maison": {"median": 1283.2, "mean": 1394.6, "q1": 833.3, "q3": 1808.8, "volume": 11335, "confidence": 0.696}, "local": {"median": 855.7, "mean": 1136.5, "q1": 500.0, "q3": 1649.5, "volume": 448, "confidence": 0.6}}, "10": {"tous": {"median": 1403.4, "mean": 1492.9, "q1": 944.7, "q3": 1904.8, "volume": 22180, "confidence": 0.726}, "appartement": {"median": 1529.7, "mean": 1649.0, "q1": 1189.2, "q3": 1948.3, "volume": 7506, "confidence": 0.802}, "maison": {"median": 1315.8, "mean": 1418.5, "q1": 844.4, "q3": 1891.0, "volume": 13764, "confidence": 0.682}, "local": {"median": 969.0, "mean": 1331.4, "q1": 559.4, "q3": 1509.4, "volume": 910, "confidence": 0.608}}, "11": {"tous": {"median": 1700.0, "mean": 1939.3, "q1": 1070.9, "q3": 2520.8, "volume": 41904, "confidence": 0.659}, "appartement": {"median": 2031.0, "mean": 2207.6, "q1": 1287.0, "q3": 2910.4, "volume": 12311, "confidence": 0.68}, "maison": {"median": 1608.6, "mean": 1852.0, "q1": 1011.9, "q3": 2355.6, "volume": 28398, "confidence": 0.666}, "local": {"median": 950.0, "mean": 1251.1, "q1": 547.8, "q3": 1571.4, "volume": 1195, "confidence": 0.6}}, "12": {"tous": {"median": 1331.3, "mean": 1435.8, "q1": 821.3, "q3": 1943.5, "volume": 21024, "confidence": 0.663}, "appartement": {"median": 1641.0, "mean": 1641.7, "q1": 1122.4, "q3": 2145.3, "volume": 6345, "confidence": 0.751}, "maison": {"median": 1223.7, "mean": 1371.3, "q1": 760.0, "q3": 1848.8, "volume": 13774, "confidence": 0.644}, "local": {"median": 755.4, "mean": 973.0, "q1": 429.4, "q3": 1235.3, "volume": 905, "confidence": 0.6}}, "13": {"tous": {"median": 3283.6, "mean": 3552.7, "q1": 2302.6, "q3": 4406.8, "volume": 149587, "confidence": 0.744}, "appartement": {"median": 3112.2, "mean": 3348.8, "q1": 2220.0, "q3": 4184.6, "volume": 96487, "confidence": 0.748}, "maison": {"median": 3798.1, "mean": 4132.5, "q1": 2830.9, "q3": 4955.2, "volume": 45523, "confidence": 0.776}, "local": {"median": 2064.3, "mean": 2665.1, "q1": 1269.2, "q3": 3271.0, "volume": 7577, "confidence": 0.612}}, "14": {"tous": {"median": 2400.0, "mean": 2644.1, "q1": 1481.5, "q3": 3381.0, "volume": 64499, "confidence": 0.683}, "appartement": {"median": 3015.9, "mean": 3312.5, "q1": 2191.0, "q3": 4105.3, "volume": 25762, "confidence": 0.746}, "maison": {"median": 1988.8, "mean": 2193.3, "q1": 1137.2, "q3": 2830.2, "volume": 36108, "confidence": 0.659}, "local": {"median": 1785.7, "mean": 2286.8, "q1": 894.6, "q3": 2947.4, "volume": 2629, "confidence": 0.6}}, "15": {"tous": {"median": 1098.6, "mean": 1216.9, "q1": 700.0, "q3": 1586.2, "volume": 10635, "confidence": 0.677}, "appartement": {"median": 1210.9, "mean": 1292.1, "q1": 883.9, "q3": 1575.5, "volume": 2845, "confidence": 0.772}, "maison": {"median": 1060.7, "mean": 1196.2, "q1": 664.7, "q3": 1612.1, "volume": 7399, "confidence": 0.643}, "local": {"median": 633.3, "mean": 1060.6, "q1": 388.3, "q3": 1044.8, "volume": 391, "confidence": 0.6}}, "16": {"tous": {"median": 1170.0, "mean": 1268.0, "q1": 757.5, "q3": 1663.0, "volume": 28466, "confidence": 0.69}, "appartement": {"median": 1486.5, "mean": 1529.6, "q1": 1156.2, "q3": 1807.0, "volume": 2841, "confidence": 0.825}, "maison": {"median": 1131.6, "mean": 1241.9, "q1": 739.6, "q3": 1641.8, "volume": 24709, "confidence": 0.681}, "local": {"median": 779.0, "mean": 1162.0, "q1": 439.7, "q3": 1294.0, "volume": 916, "confidence": 0.6}}, "17": {"tous": {"median": 2538.5, "mean": 2917.4, "q1": 1501.8, "q3": 3898.6, "volume": 68098, "confidence": 0.622}, "appartement": {"median": 3692.3, "mean": 3784.1, "q1": 2398.4, "q3": 4963.0, "volume": 14112, "confidence": 0.722}, "maison": {"median": 2307.7, "mean": 2706.4, "q1": 1358.0, "q3": 3542.2, "volume": 51383, "confidence": 0.621}, "local": {"median": 1817.2, "mean": 2383.2, "q1": 945.9, "q3": 3302.3, "volume": 2603, "confidence": 0.6}}, "18": {"tous": {"median": 1000.0, "mean": 1128.5, "q1": 647.5, "q3": 1489.8, "volume": 24469, "confidence": 0.663}, "appartement": {"median": 1282.0, "mean": 1312.1, "q1": 865.4, "q3": 1698.1, "volume": 3982, "confidence": 0.74}, "maison": {"median": 958.3, "mean": 1098.8, "q1": 626.9, "q3": 1433.7, "volume": 19736, "confidence": 0.663}, "local": {"median": 700.0, "mean": 937.1, "q1": 400.0, "q3": 1150.4, "volume": 751, "confidence": 0.6}}, "19": {"tous": {"median": 1132.4, "mean": 1261.2, "q1": 721.2, "q3": 1666.7, "volume": 18577, "confidence": 0.666}, "appartement": {"median": 1300.0, "mean": 1360.8, "q1": 873.0, "q3": 1755.6, "volume": 4063, "confidence": 0.728}, "maison": {"median": 1094.6, "mean": 1248.8, "q1": 714.3, "q3": 1650.9, "volume": 13885, "confidence": 0.658}, "local": {"median": 666.7, "mean": 892.0, "q1": 407.6, "q3": 1100.0, "volume": 629, "confidence": 0.6}}, "21": {"tous": {"median": 2166.7, "mean": 2181.1, "q1": 1371.8, "q3": 2833.3, "volume": 43798, "confidence": 0.73}, "appartement": {"median": 2416.7, "mean": 2417.0, "q1": 1831.3, "q3": 2971.4, "volume": 20913, "confidence": 0.811}, "maison": {"median": 1852.6, "mean": 1970.9, "q1": 1076.4, "q3": 2666.7, "volume": 20961, "confidence": 0.657}, "local": {"median": 1500.0, "mean": 1906.1, "q1": 774.2, "q3": 2368.4, "volume": 1924, "confidence": 0.6}}, "22": {"tous": {"median": 1574.6, "mean": 1793.6, "q1": 967.7, "q3": 2311.8, "volume": 55291, "confidence": 0.659}, "appartement": {"median": 1958.3, "mean": 2244.5, "q1": 1426.8, "q3": 2758.6, "volume": 9533, "confidence": 0.728}, "maison": {"median": 1500.0, "mean": 1711.2, "q1": 911.5, "q3": 2226.1, "volume": 43708, "confidence": 0.649}, "local": {"median": 1027.4, "mean": 1453.6, "q1": 575.5, "q3": 1866.7, "volume": 2050, "confidence": 0.6}}, "23": {"tous": {"median": 710.9, "mean": 831.5, "q1": 461.5, "q3": 1050.0, "volume": 10623, "confidence": 0.669}, "appartement": {"median": 714.3, "mean": 761.4, "q1": 486.9, "q3": 971.8, "volume": 752, "confidence": 0.728}, "maison": {"median": 714.3, "mean": 838.7, "q1": 463.9, "q3": 1060.6, "volume": 9687, "confidence": 0.666}, "local": {"median": 473.1, "mean": 740.0, "q1": 326.1, "q3": 833.3, "volume": 184, "confidence": 0.6}}, "24": {"tous": {"median": 1214.3, "mean": 1359.1, "q1": 808.1, "q3": 1727.3, "volume": 36883, "confidence": 0.697}, "appartement": {"median": 1478.4, "mean": 1527.9, "q1": 1162.2, "q3": 1822.6, "volume": 4927, "confidence": 0.821}, "maison": {"median": 1166.7, "mean": 1330.9, "q1": 784.7, "q3": 1705.7, "volume": 30766, "confidence": 0.684}, "local": {"median": 854.5, "mean": 1389.3, "q1": 488.9, "q3": 1486.5, "volume": 1190, "confidence": 0.6}}, "25": {"tous": {"median": 1830.2, "mean": 1916.2, "q1": 1194.6, "q3": 2500.0, "volume": 38158, "confidence": 0.715}, "appartement": {"median": 1906.2, "mean": 1953.0, "q1": 1290.9, "q3": 2528.6, "volume": 20055, "confidence": 0.74}, "maison": {"median": 1800.0, "mean": 1927.4, "q1": 1180.8, "q3": 2509.3, "volume": 16302, "confidence": 0.705}, "local": {"median": 1136.4, "mean": 1406.0, "q1": 615.4, "q3": 1796.4, "volume": 1801, "confidence": 0.6}}, "26": {"tous": {"median": 1927.5, "mean": 2016.8, "q1": 1366.3, "q3": 2551.0, "volume": 37930, "confidence": 0.754}, "appartement": {"median": 1760.0, "mean": 1841.8, "q1": 1338.7, "q3": 2263.2, "volume": 13998, "confidence": 0.79}, "maison": {"median": 2133.3, "mean": 2170.1, "q1": 1478.7, "q3": 2714.3, "volume": 21876, "confidence": 0.768}, "local": {"median": 1192.6, "mean": 1576.9, "q1": 703.1, "q3": 1921.8, "volume": 2056, "confidence": 0.6}}, "27": {"tous": {"median": 1515.5, "mean": 1613.4, "q1": 1006.4, "q3": 2084.3, "volume": 42426, "confidence": 0.716}, "appartement": {"median": 1746.3, "mean": 1812.5, "q1": 1372.5, "q3": 2197.4, "volume": 5919, "confidence": 0.811}, "maison": {"median": 1470.2, "mean": 1590.2, "q1": 977.3, "q3": 2071.4, "volume": 35334, "confidence": 0.702}, "local": {"median": 960.7, "mean": 1309.6, "q1": 535.7, "q3": 1625.0, "volume": 1173, "confidence": 0.6}}, "28": {"tous": {"median": 1655.6, "mean": 1742.6, "q1": 1075.9, "q3": 2275.6, "volume": 31209, "confidence": 0.71}, "appartement": {"median": 2059.5, "mean": 2043.8, "q1": 1436.4, "q3": 2600.0, "volume": 6087, "confidence": 0.774}, "maison": {"median": 1575.8, "mean": 1674.2, "q1": 1026.5, "q3": 2182.7, "volume": 24261, "confidence": 0.707}, "local": {"median": 994.9, "mean": 1539.4, "q1": 555.6, "q3": 1750.0, "volume": 861, "confidence": 0.6}}, "29": {"tous": {"median": 1818.2, "mean": 1924.9, "q1": 1231.9, "q3": 2432.4, "volume": 78321, "confidence": 0.736}, "appartement": {"median": 1971.8, "mean": 2094.8, "q1": 1529.8, "q3": 2515.0, "volume": 23098, "confidence": 0.8}, "maison": {"median": 1758.9, "mean": 1885.9, "q1": 1148.5, "q3": 2420.0, "volume": 52020, "confidence": 0.711}, "local": {"median": 1103.4, "mean": 1332.8, "q1": 631.6, "q3": 1797.9, "volume": 3203, "confidence": 0.6}}, "2A": {"tous": {"median": 3580.2, "mean": 3935.7, "q1": 2414.2, "q3": 4993.6, "volume": 11405, "confidence": 0.712}, "appartement": {"median": 3623.1, "mean": 3796.1, "q1": 2608.7, "q3": 4821.4, "volume": 6834, "confidence": 0.756}, "maison": {"median": 3812.6, "mean": 4459.6, "q1": 2271.9, "q3": 5700.0, "volume": 3746, "confidence": 0.64}, "local": {"median": 2411.8, "mean": 2713.4, "q1": 1428.6, "q3": 3555.6, "volume": 825, "confidence": 0.647}}, "2B": {"tous": {"median": 2894.7, "mean": 3047.9, "q1": 2000.0, "q3": 3800.0, "volume": 11685, "confidence": 0.751}, "appartement": {"median": 2924.5, "mean": 3051.2, "q1": 2131.9, "q3": 3777.8, "volume": 7342, "confidence": 0.775}, "maison": {"median": 2963.2, "mean": 3179.1, "q1": 1859.0, "q3": 3966.5, "volume": 3753, "confidence": 0.716}, "local": {"median": 1916.7, "mean": 2171.6, "q1": 1048.2, "q3": 2777.8, "volume": 590, "confidence": 0.639}}, "30": {"tous": {"median": 2176.1, "mean": 2382.8, "q1": 1415.4, "q3": 2994.0, "volume": 61106, "confidence": 0.71}, "appartement": {"median": 2066.7, "mean": 2448.3, "q1": 1325.0, "q3": 3075.0, "volume": 23317, "confidence": 0.661}, "maison": {"median": 2297.3, "mean": 2399.3, "q1": 1548.7, "q3": 3000.0, "volume": 35239, "confidence": 0.747}, "local": {"median": 1286.5, "mean": 1556.1, "q1": 760.0, "q3": 2000.0, "volume": 2550, "confidence": 0.614}}, "31": {"tous": {"median": 2718.4, "mean": 2828.5, "q1": 1972.2, "q3": 3468.1, "volume": 105201, "confidence": 0.78}, "appartement": {"median": 2875.0, "mean": 3069.4, "q1": 2295.7, "q3": 3673.8, "volume": 54728, "confidence": 0.808}, "maison": {"median": 2543.9, "mean": 2585.7, "q1": 1588.2, "q3": 3263.7, "volume": 46518, "confidence": 0.737}, "local": {"median": 1882.0, "mean": 2351.1, "q1": 1153.8, "q3": 2925.0, "volume": 3955, "confidence": 0.624}}, "32": {"tous": {"median": 1294.0, "mean": 1412.3, "q1": 873.2, "q3": 1804.6, "volume": 14858, "confidence": 0.712}, "appartement": {"median": 1445.3, "mean": 1473.4, "q1": 1125.0, "q3": 1779.2, "volume": 2273, "confidence": 0.819}, "maison": {"median": 1275.2, "mean": 1419.0, "q1": 854.2, "q3": 1830.1, "volume": 12124, "confidence": 0.694}, "local": {"median": 714.3, "mean": 935.8, "q1": 435.7, "q3": 1134.0, "volume": 461, "confidence": 0.609}}, "33": {"tous": {"median": 3351.1, "mean": 3543.5, "q1": 2126.3, "q3": 4464.3, "volume": 126310, "confidence": 0.721}, "appartement": {"median": 3776.3, "mean": 4001.3, "q1": 2944.4, "q3": 4736.1, "volume": 47692, "confidence": 0.81}, "maison": {"median": 2976.2, "mean": 3286.0, "q1": 1717.6, "q3": 4268.3, "volume": 74176, "confidence": 0.657}, "local": {"median": 2299.4, "mean": 2929.6, "q1": 1321.2, "q3": 3788.4, "volume": 4442, "confidence": 0.6}}, "34": {"tous": {"median": 2940.7, "mean": 3074.1, "q1": 1944.4, "q3": 3948.3, "volume": 118740, "confidence": 0.727}, "appartement": {"median": 3205.9, "mean": 3315.5, "q1": 2232.1, "q3": 4166.7, "volume": 63508, "confidence": 0.759}, "maison": {"median": 2698.8, "mean": 2857.4, "q1": 1751.3, "q3": 3705.1, "volume": 49976, "confidence": 0.71}, "local": {"median": 1902.8, "mean": 2216.6, "q1": 1145.0, "q3": 2875.4, "volume": 5256, "confidence": 0.636}}, "35": {"tous": {"median": 2562.5, "mean": 2768.9, "q1": 1688.2, "q3": 3550.0, "volume": 80001, "confidence": 0.709}, "appartement": {"median": 3238.7, "mean": 3415.0, "q1": 2500.0, "q3": 4156.2, "volume": 31218, "confidence": 0.795}, "maison": {"median": 2117.6, "mean": 2379.2, "q1": 1321.8, "q3": 2961.5, "volume": 45505, "confidence": 0.69}, "local": {"median": 1598.7, "mean": 2026.1, "q1": 914.3, "q3": 2661.3, "volume": 3278, "confidence": 0.6}}, "36": {"tous": {"median": 895.5, "mean": 1003.7, "q1": 589.3, "q3": 1289.5, "volume": 17022, "confidence": 0.687}, "appartement": {"median": 1026.4, "mean": 1059.4, "q1": 771.9, "q3": 1278.5, "volume": 1613, "confidence": 0.803}, "maison": {"median": 884.1, "mean": 999.5, "q1": 584.0, "q3": 1294.1, "volume": 14964, "confidence": 0.679}, "local": {"median": 598.5, "mean": 945.6, "q1": 370.4, "q3": 1031.2, "volume": 445, "confidence": 0.6}}, "37": {"tous": {"median": 2069.5, "mean": 2170.6, "q1": 1291.4, "q3": 2844.4, "volume": 45530, "confidence": 0.7}, "appartement": {"median": 2579.9, "mean": 2586.9, "q1": 1899.0, "q3": 3179.5, "volume": 14294, "confidence": 0.801}, "maison": {"median": 1822.9, "mean": 1996.7, "q1": 1117.6, "q3": 2628.6, "volume": 29305, "confidence": 0.668}, "local": {"median": 1241.7, "mean": 1727.5, "q1": 681.2, "q3": 2173.9, "volume": 1931, "confidence": 0.6}}, "38": {"tous": {"median": 2420.1, "mean": 2571.7, "q1": 1724.5, "q3": 3197.7, "volume": 93703, "confidence": 0.757}, "appartement": {"median": 2443.8, "mean": 2628.5, "q1": 1839.5, "q3": 3166.4, "volume": 49243, "confidence": 0.783}, "maison": {"median": 2500.0, "mean": 2612.3, "q1": 1700.4, "q3": 3309.9, "volume": 39048, "confidence": 0.742}, "local": {"median": 1408.5, "mean": 1762.5, "q1": 882.4, "q3": 2255.7, "volume": 5412, "confidence": 0.61}}, "39": {"tous": {"median": 1342.8, "mean": 1485.6, "q1": 864.7, "q3": 1902.7, "volume": 17911, "confidence": 0.691}, "appartement": {"median": 1327.4, "mean": 1487.3, "q1": 911.9, "q3": 1834.6, "volume": 5888, "confidence": 0.722}, "maison": {"median": 1400.0, "mean": 1516.9, "q1": 884.6, "q3": 1959.6, "volume": 11266, "confidence": 0.693}, "local": {"median": 719.3, "mean": 1006.4, "q1": 408.2, "q3": 1220.3, "volume": 757, "confidence": 0.6}}, "40": {"tous": {"median": 2359.1, "mean": 2836.0, "q1": 1553.8, "q3": 3565.2, "volume": 38674, "confidence": 0.659}, "appartement": {"median": 2851.9, "mean": 3379.5, "q1": 2083.3, "q3": 4266.7, "volume": 11301, "confidence": 0.694}, "maison": {"median": 2195.1, "mean": 2661.5, "q1": 1386.8, "q3": 3333.3, "volume": 25876, "confidence": 0.645}, "local": {"median": 1486.5, "mean": 1748.3, "q1": 803.1, "q3": 2410.0, "volume": 1497, "confidence": 0.6}}, "41": {"tous": {"median": 1250.0, "mean": 1360.6, "q1": 813.3, "q3": 1764.7, "volume": 25118, "confidence": 0.696}, "appartement": {"median": 1646.1, "mean": 1628.2, "q1": 1193.8, "q3": 2023.8, "volume": 3583, "confidence": 0.798}, "maison": {"median": 1194.4, "mean": 1320.2, "q1": 791.1, "q3": 1703.7, "volume": 20577, "confidence": 0.694}, "local": {"median": 844.8, "mean": 1228.8, "q1": 466.7, "q3": 1400.0, "volume": 958, "confidence": 0.6}}, "42": {"tous": {"median": 1333.3, "mean": 1509.5, "q1": 949.0, "q3": 1935.5, "volume": 55648, "confidence": 0.704}, "appartement": {"median": 1214.3, "mean": 1321.8, "q1": 925.0, "q3": 1592.1, "volume": 27539, "confidence": 0.78}, "maison": {"median": 1674.9, "mean": 1778.4, "q1": 1094.3, "q3": 2347.0, "volume": 25044, "confidence": 0.701}, "local": {"median": 795.5, "mean": 999.7, "q1": 483.9, "q3": 1220.3, "volume": 3065, "confidence": 0.63}}, "43": {"tous": {"median": 1159.1, "mean": 1282.4, "q1": 756.1, "q3": 1688.2, "volume": 15789, "confidence": 0.678}, "appartement": {"median": 1227.3, "mean": 1289.2, "q1": 898.9, "q3": 1636.4, "volume": 3591, "confidence": 0.76}, "maison": {"median": 1160.7, "mean": 1304.4, "q1": 750.0, "q3": 1740.0, "volume": 11553, "confidence": 0.659}, "local": {"median": 633.6, "mean": 850.0, "q1": 393.4, "q3": 1034.5, "volume": 645, "confidence": 0.6}}, "44": {"tous": {"median": 3030.3, "mean": 3188.0, "q1": 2086.5, "q3": 3962.3, "volume": 113530, "confidence": 0.752}, "appartement": {"median": 3475.7, "mean": 3679.6, "q1": 2693.9, "q3": 4328.4, "volume": 42694, "confidence": 0.812}, "maison": {"median": 2750.0, "mean": 2937.6, "q1": 1776.9, "q3": 3695.7, "volume": 66161, "confidence": 0.721}, "local": {"median": 1826.5, "mean": 2241.7, "q1": 1033.6, "q3": 2894.7, "volume": 4675, "confidence": 0.6}}, "45": {"tous": {"median": 1692.3, "mean": 1808.8, "q1": 1088.9, "q3": 2348.5, "volume": 48866, "confidence": 0.702}, "appartement": {"median": 2118.6, "mean": 2176.1, "q1": 1553.2, "q3": 2750.5, "volume": 13805, "confidence": 0.774}, "maison": {"median": 1524.3, "mean": 1654.4, "q1": 983.5, "q3": 2157.6, "volume": 33253, "confidence": 0.692}, "local": {"median": 1219.8, "mean": 1844.8, "q1": 648.0, "q3": 2164.3, "volume": 1808, "confidence": 0.6}}, "46": {"tous": {"median": 1234.5, "mean": 1361.3, "q1": 844.7, "q3": 1740.0, "volume": 14418, "confidence": 0.71}, "appartement": {"median": 1333.3, "mean": 1346.5, "q1": 997.1, "q3": 1666.7, "volume": 1923, "confidence": 0.799}, "maison": {"median": 1232.5, "mean": 1380.8, "q1": 846.4, "q3": 1775.7, "volume": 11956, "confidence": 0.698}, "local": {"median": 764.9, "mean": 980.6, "q1": 455.3, "q3": 1237.0, "volume": 539, "confidence": 0.6}}, "47": {"tous": {"median": 1196.1, "mean": 1322.2, "q1": 797.2, "q3": 1701.4, "volume": 27456, "confidence": 0.698}, "appartement": {"median": 1359.9, "mean": 1402.5, "q1": 1000.0, "q3": 1721.9, "volume": 4764, "confidence": 0.788}, "maison": {"median": 1166.7, "mean": 1299.4, "q1": 778.3, "q3": 1702.1, "volume": 21627, "confidence": 0.683}, "local": {"median": 784.3, "mean": 1425.4, "q1": 465.3, "q3": 1428.6, "volume": 1065, "confidence": 0.6}}, "48": {"tous": {"median": 1250.0, "mean": 1382.7, "q1": 810.0, "q3": 1820.4, "volume": 4885, "confidence": 0.677}, "appartement": {"median": 1195.1, "mean": 1270.7, "q1": 789.5, "q3": 1689.6, "volume": 1027, "confidence": 0.699}, "maison": {"median": 1301.4, "mean": 1437.7, "q1": 846.2, "q3": 1863.6, "volume": 3691, "confidence": 0.687}, "local": {"median": 646.0, "mean": 855.3, "q1": 388.7, "q3": 1058.8, "volume": 167, "confidence": 0.6}}, "49": {"tous": {"median": 1936.9, "mean": 2076.8, "q1": 1242.4, "q3": 2678.6, "volume": 56014, "confidence": 0.703}, "appartement": {"median": 2625.6, "mean": 2705.0, "q1": 1895.2, "q3": 3421.1, "volume": 14293, "confidence": 0.768}, "maison": {"median": 1761.4, "mean": 1866.7, "q1": 1136.4, "q3": 2398.6, "volume": 39175, "confidence": 0.713}, "local": {"median": 1139.1, "mean": 1782.6, "q1": 596.2, "q3": 2227.3, "volume": 2546, "confidence": 0.6}}, "50": {"tous": {"median": 1500.0, "mean": 1714.5, "q1": 896.9, "q3": 2234.5, "volume": 39558, "confidence": 0.643}, "appartement": {"median": 1964.3, "mean": 2239.8, "q1": 1431.6, "q3": 2769.2, "volume": 5510, "confidence": 0.728}, "maison": {"median": 1422.8, "mean": 1643.6, "q1": 852.0, "q3": 2162.8, "volume": 32803, "confidence": 0.631}, "local": {"median": 942.6, "mean": 1259.3, "q1": 535.7, "q3": 1597.2, "volume": 1245, "confidence": 0.6}}, "51": {"tous": {"median": 1933.3, "mean": 2047.2, "q1": 1304.3, "q3": 2625.0, "volume": 37107, "confidence": 0.727}, "appartement": {"median": 2184.6, "mean": 2285.0, "q1": 1590.9, "q3": 2869.0, "volume": 15037, "confidence": 0.766}, "maison": {"median": 1779.5, "mean": 1887.6, "q1": 1162.8, "q3": 2422.7, "volume": 20414, "confidence": 0.717}, "local": {"median": 1272.2, "mean": 1856.1, "q1": 722.2, "q3": 2352.9, "volume": 1656, "confidence": 0.6}}, "52": {"tous": {"median": 887.0, "mean": 968.8, "q1": 535.7, "q3": 1312.5, "volume": 10684, "confidence": 0.65}, "appartement": {"median": 1010.8, "mean": 1010.0, "q1": 763.6, "q3": 1236.6, "volume": 1364, "confidence": 0.813}, "maison": {"median": 860.2, "mean": 970.4, "q1": 517.7, "q3": 1343.8, "volume": 8989, "confidence": 0.616}, "local": {"median": 625.0, "mean": 757.2, "q1": 353.7, "q3": 1000.0, "volume": 331, "confidence": 0.6}}, "53": {"tous": {"median": 1296.3, "mean": 1376.1, "q1": 790.0, "q3": 1851.5, "volume": 23269, "confidence": 0.672}, "appartement": {"median": 1698.1, "mean": 1724.4, "q1": 1242.4, "q3": 2150.5, "volume": 3026, "confidence": 0.786}, "maison": {"median": 1243.1, "mean": 1330.3, "q1": 759.5, "q3": 1796.6, "volume": 19550, "confidence": 0.666}, "local": {"median": 800.0, "mean": 1149.0, "q1": 463.9, "q3": 1320.0, "volume": 693, "confidence": 0.6}}, "54": {"tous": {"median": 1733.8, "mean": 1784.4, "q1": 1145.8, "q3": 2314.3, "volume": 52852, "confidence": 0.73}, "appartement": {"median": 1803.6, "mean": 1829.3, "q1": 1258.8, "q3": 2327.5, "volume": 22396, "confidence": 0.763}, "maison": {"median": 1708.8, "mean": 1770.6, "q1": 1100.0, "q3": 2321.0, "volume": 28474, "confidence": 0.714}, "local": {"median": 1129.3, "mean": 1475.6, "q1": 650.0, "q3": 1842.1, "volume": 1982, "confidence": 0.6}}, "55": {"tous": {"median": 905.3, "mean": 1005.5, "q1": 576.9, "q3": 1333.3, "volume": 11660, "confidence": 0.666}, "appartement": {"median": 949.7, "mean": 989.6, "q1": 656.6, "q3": 1240.7, "volume": 1636, "confidence": 0.754}, "maison": {"median": 900.0, "mean": 1010.3, "q1": 568.4, "q3": 1355.6, "volume": 9736, "confidence": 0.65}, "local": {"median": 690.2, "mean": 931.1, "q1": 404.2, "q3": 1142.9, "volume": 288, "confidence": 0.6}}, "56": {"tous": {"median": 2264.4, "mean": 2529.1, "q1": 1300.0, "q3": 3375.0, "volume": 66390, "confidence": 0.633}, "appartement": {"median": 3012.3, "mean": 3213.2, "q1": 2152.5, "q3": 3981.5, "volume": 16777, "confidence": 0.757}, "maison": {"median": 2024.2, "mean": 2336.4, "q1": 1128.7, "q3": 3113.4, "volume": 46761, "confidence": 0.608}, "local": {"median": 1300.6, "mean": 1665.7, "q1": 733.3, "q3": 2200.0, "volume": 2852, "confidence": 0.6}}, "58": {"tous": {"median": 888.9, "mean": 1003.4, "q1": 578.9, "q3": 1277.8, "volume": 18420, "confidence": 0.686}, "appartement": {"median": 896.3, "mean": 923.1, "q1": 629.1, "q3": 1155.2, "volume": 2858, "confidence": 0.765}, "maison": {"median": 892.9, "mean": 1013.8, "q1": 579.7, "q3": 1318.0, "volume": 15040, "confidence": 0.669}, "local": {"median": 600.0, "mean": 1143.6, "q1": 359.0, "q3": 1060.6, "volume": 522, "confidence": 0.6}}, "59": {"tous": {"median": 1950.0, "mean": 2177.8, "q1": 1273.5, "q3": 2833.3, "volume": 168756, "confidence": 0.68}, "appartement": {"median": 2617.2, "mean": 2789.7, "q1": 1753.6, "q3": 3630.8, "volume": 44138, "confidence": 0.713}, "maison": {"median": 1788.1, "mean": 1962.1, "q1": 1184.2, "q3": 2531.9, "volume": 118578, "confidence": 0.699}, "local": {"median": 1305.8, "mean": 1940.1, "q1": 701.8, "q3": 2457.6, "volume": 6040, "confidence": 0.6}}, "60": {"tous": {"median": 2075.0, "mean": 2207.1, "q1": 1456.5, "q3": 2753.0, "volume": 52270, "confidence": 0.75}, "appartement": {"median": 2200.0, "mean": 2360.9, "q1": 1580.0, "q3": 2985.1, "volume": 12578, "confidence": 0.745}, "maison": {"median": 2058.8, "mean": 2162.1, "q1": 1437.5, "q3": 2691.5, "volume": 38044, "confidence": 0.756}, "local": {"median": 1475.4, "mean": 2070.4, "q1": 803.6, "q3": 2540.8, "volume": 1648, "confidence": 0.6}}, "61": {"tous": {"median": 1015.0, "mean": 1140.1, "q1": 655.0, "q3": 1470.0, "volume": 21482, "confidence": 0.679}, "appartement": {"median": 1155.6, "mean": 1250.9, "q1": 846.2, "q3": 1566.0, "volume": 2366, "confidence": 0.751}, "maison": {"median": 1000.0, "mean": 1136.1, "q1": 650.0, "q3": 1467.5, "volume": 18455, "confidence": 0.673}, "local": {"median": 587.8, "mean": 854.5, "q1": 357.1, "q3": 973.5, "volume": 661, "confidence": 0.6}}, "62": {"tous": {"median": 1569.3, "mean": 1862.4, "q1": 1098.9, "q3": 2125.2, "volume": 89090, "confidence": 0.738}, "appartement": {"median": 2074.1, "mean": 2842.0, "q1": 1410.7, "q3": 3333.3, "volume": 15170, "confidence": 0.629}, "maison": {"median": 1516.1, "mean": 1664.9, "q1": 1070.8, "q3": 2000.0, "volume": 71120, "confidence": 0.755}, "local": {"median": 978.6, "mean": 1569.8, "q1": 555.6, "q3": 1737.9, "volume": 2800, "confidence": 0.6}}, "63": {"tous": {"median": 1734.7, "mean": 1777.4, "q1": 1052.2, "q3": 2371.1, "volume": 48373, "confidence": 0.696}, "appartement": {"median": 2011.3, "mean": 2048.6, "q1": 1550.0, "q3": 2512.2, "volume": 17759, "confidence": 0.809}, "maison": {"median": 1484.5, "mean": 1639.8, "q1": 866.7, "q3": 2272.7, "volume": 28830, "confidence": 0.621}, "local": {"median": 1022.5, "mean": 1300.8, "q1": 591.1, "q3": 1696.4, "volume": 1784, "confidence": 0.6}}, "64": {"tous": {"median": 2444.4, "mean": 3210.6, "q1": 1574.1, "q3": 4230.8, "volume": 57491, "confidence": 0.6}, "appartement": {"median": 3068.0, "mean": 3726.5, "q1": 1942.9, "q3": 4863.9, "volume": 31645, "confidence": 0.619}, "maison": {"median": 2020.0, "mean": 2642.7, "q1": 1229.5, "q3": 3082.8, "volume": 22995, "confidence": 0.633}, "local": {"median": 1513.3, "mean": 2064.8, "q1": 819.0, "q3": 2546.5, "volume": 2851, "confidence": 0.6}}, "65": {"tous": {"median": 1495.3, "mean": 1704.3, "q1": 1036.6, "q3": 2122.0, "volume": 21988, "confidence": 0.71}, "appartement": {"median": 1550.0, "mean": 1860.7, "q1": 1111.1, "q3": 2450.0, "volume": 9970, "confidence": 0.654}, "maison": {"median": 1497.5, "mean": 1594.0, "q1": 1013.6, "q3": 2000.0, "volume": 11159, "confidence": 0.737}, "local": {"median": 763.6, "mean": 1322.8, "q1": 448.2, "q3": 1250.0, "volume": 859, "confidence": 0.6}}, "66": {"tous": {"median": 2168.7, "mean": 2349.5, "q1": 1421.9, "q3": 3019.9, "volume": 60224, "confidence": 0.705}, "appartement": {"median": 2073.2, "mean": 2341.5, "q1": 1358.8, "q3": 3095.2, "volume": 27223, "confidence": 0.665}, "maison": {"median": 2295.6, "mean": 2416.5, "q1": 1573.1, "q3": 3013.9, "volume": 30564, "confidence": 0.749}, "local": {"median": 1333.3, "mean": 1597.0, "q1": 791.7, "q3": 2076.9, "volume": 2437, "confidence": 0.614}}, "69": {"tous": {"median": 3651.7, "mean": 3743.9, "q1": 2547.6, "q3": 4733.3, "volume": 123788, "confidence": 0.761}, "appartement": {"median": 3846.2, "mean": 3933.0, "q1": 2839.2, "q3": 4889.7, "volume": 84201, "confidence": 0.787}, "maison": {"median": 3333.3, "mean": 3488.7, "q1": 2154.8, "q3": 4436.6, "volume": 32288, "confidence": 0.726}, "local": {"median": 2077.5, "mean": 2692.4, "q1": 1210.7, "q3": 3481.2, "volume": 7299, "confidence": 0.6}}, "70": {"tous": {"median": 972.8, "mean": 1083.9, "q1": 634.9, "q3": 1401.6, "volume": 14482, "confidence": 0.685}, "appartement": {"median": 950.0, "mean": 989.4, "q1": 678.6, "q3": 1225.8, "volume": 2695, "confidence": 0.77}, "maison": {"median": 1000.0, "mean": 1119.1, "q1": 645.2, "q3": 1478.8, "volume": 11312, "confidence": 0.667}, "local": {"median": 615.4, "mean": 782.9, "q1": 382.7, "q3": 978.6, "volume": 475, "confidence": 0.613}}, "71": {"tous": {"median": 1180.6, "mean": 1316.5, "q1": 783.1, "q3": 1707.3, "volume": 40251, "confidence": 0.687}, "appartement": {"median": 1191.2, "mean": 1286.3, "q1": 857.1, "q3": 1616.6, "volume": 10310, "confidence": 0.745}, "maison": {"median": 1196.6, "mean": 1339.9, "q1": 781.2, "q3": 1763.7, "volume": 28356, "confidence": 0.672}, "local": {"median": 737.2, "mean": 1094.2, "q1": 434.8, "q3": 1266.7, "volume": 1585, "confidence": 0.6}}, "72": {"tous": {"median": 1366.3, "mean": 1443.3, "q1": 864.3, "q3": 1916.7, "volume": 42319, "confidence": 0.692}, "appartement": {"median": 1671.5, "mean": 1739.4, "q1": 1259.3, "q3": 2131.1, "volume": 7772, "confidence": 0.791}, "maison": {"median": 1290.9, "mean": 1386.7, "q1": 809.5, "q3": 1860.5, "volume": 33115, "confidence": 0.674}, "local": {"median": 862.9, "mean": 1145.1, "q1": 488.4, "q3": 1396.6, "volume": 1432, "confidence": 0.6}}, "73": {"tous": {"median": 3159.2, "mean": 3778.2, "q1": 2142.9, "q3": 4500.0, "volume": 47162, "confidence": 0.702}, "appartement": {"median": 3399.9, "mean": 4094.0, "q1": 2414.9, "q3": 4778.6, "volume": 31057, "confidence": 0.722}, "maison": {"median": 2745.0, "mean": 3178.9, "q1": 1787.1, "q3": 3951.9, "volume": 13000, "confidence": 0.685}, "local": {"median": 2138.9, "mean": 3127.9, "q1": 1156.6, "q3": 4027.8, "volume": 3105, "confidence": 0.6}}, "74": {"tous": {"median": 4000.0, "mean": 4407.2, "q1": 2979.6, "q3": 5240.8, "volume": 77027, "confidence": 0.774}, "appartement": {"median": 4023.8, "mean": 4411.1, "q1": 3080.4, "q3": 5227.3, "volume": 51353, "confidence": 0.787}, "maison": {"median": 4137.0, "mean": 4624.3, "q1": 2981.5, "q3": 5429.3, "volume": 22456, "confidence": 0.763}, "local": {"median": 2307.7, "mean": 2830.7, "q1": 1351.7, "q3": 3571.4, "volume": 3218, "confidence": 0.615}}, "75": {"tous": {"median": 10284.5, "mean": 10351.6, "q1": 8541.6, "q3": 12083.3, "volume": 159360, "confidence": 0.862}, "appartement": {"median": 10362.8, "mean": 10477.6, "q1": 8709.5, "q3": 12106.8, "volume": 147088, "confidence": 0.869}, "maison": {"median": 13457.8, "mean": 13777.4, "q1": 10400.0, "q3": 17260.3, "volume": 652, "confidence": 0.796}, "local": {"median": 8255.6, "mean": 8564.9, "q1": 5411.0, "q3": 11215.0, "volume": 11620, "confidence": 0.719}}, "76": {"tous": {"median": 1900.0, "mean": 2007.4, "q1": 1269.8, "q3": 2555.3, "volume": 90674, "confidence": 0.729}, "appartement": {"median": 2153.8, "mean": 2257.1, "q1": 1534.0, "q3": 2864.8, "volume": 32914, "confidence": 0.753}, "maison": {"median": 1781.2, "mean": 1870.3, "q1": 1160.7, "q3": 2376.5, "volume": 54432, "confidence": 0.727}, "local": {"median": 1285.2, "mean": 1779.6, "q1": 732.9, "q3": 2232.0, "volume": 3328, "confidence": 0.6}}, "77": {"tous": {"median": 2867.2, "mean": 2913.1, "q1": 1954.5, "q3": 3692.3, "volume": 91321, "confidence": 0.758}, "appartement": {"median": 3228.3, "mean": 3259.3, "q1": 2473.4, "q3": 3979.6, "volume": 33952, "confidence": 0.813}, "maison": {"median": 2641.8, "mean": 2725.9, "q1": 1731.6, "q3": 3485.7, "volume": 53910, "confidence": 0.734}, "local": {"median": 1800.0, "mean": 2431.2, "q1": 1083.3, "q3": 2987.0, "volume": 3459, "confidence": 0.6}}, "78": {"tous": {"median": 3790.0, "mean": 4155.8, "q1": 2783.7, "q3": 5081.6, "volume": 88078, "confidence": 0.757}, "appartement": {"median": 3966.7, "mean": 4284.9, "q1": 3000.0, "q3": 5204.1, "volume": 45616, "confidence": 0.778}, "maison": {"median": 3636.4, "mean": 4091.1, "q1": 2607.4, "q3": 4983.3, "volume": 39315, "confidence": 0.739}, "local": {"median": 2397.3, "mean": 3094.8, "q1": 1363.6, "q3": 4038.5, "volume": 3147, "confidence": 0.6}}, "79": {"tous": {"median": 1167.0, "mean": 1279.0, "q1": 750.0, "q3": 1700.0, "volume": 31041, "confidence": 0.674}, "appartement": {"median": 1604.7, "mean": 1644.8, "q1": 1149.1, "q3": 2055.6, "volume": 2840, "confidence": 0.774}, "maison": {"median": 1134.2, "mean": 1246.2, "q1": 729.9, "q3": 1658.3, "volume": 27200, "confidence": 0.673}, "local": {"median": 833.3, "mean": 1132.6, "q1": 458.3, "q3": 1428.6, "volume": 1001, "confidence": 0.6}}, "80": {"tous": {"median": 1600.0, "mean": 1784.1, "q1": 1029.0, "q3": 2307.7, "volume": 38817, "confidence": 0.68}, "appartement": {"median": 2326.5, "mean": 2454.1, "q1": 1698.7, "q3": 2996.2, "volume": 7619, "confidence": 0.777}, "maison": {"median": 1461.3, "mean": 1621.5, "q1": 952.8, "q3": 2083.3, "volume": 30105, "confidence": 0.691}, "local": {"median": 1191.4, "mean": 1592.6, "q1": 660.0, "q3": 2000.0, "volume": 1093, "confidence": 0.6}}, "81": {"tous": {"median": 1362.5, "mean": 1491.6, "q1": 906.0, "q3": 1961.5, "volume": 26542, "confidence": 0.69}, "appartement": {"median": 1710.9, "mean": 1748.7, "q1": 1203.7, "q3": 2187.5, "volume": 4300, "confidence": 0.77}, "maison": {"median": 1306.9, "mean": 1454.8, "q1": 888.8, "q3": 1910.7, "volume": 21376, "confidence": 0.687}, "local": {"median": 750.0, "mean": 1124.6, "q1": 436.0, "q3": 1352.9, "volume": 866, "confidence": 0.6}}, "82": {"tous": {"median": 1460.7, "mean": 1541.5, "q1": 975.6, "q3": 2002.5, "volume": 18714, "confidence": 0.719}, "appartement": {"median": 1636.4, "mean": 1647.4, "q1": 1303.6, "q3": 2000.0, "volume": 3733, "confidence": 0.83}, "maison": {"median": 1409.5, "mean": 1528.2, "q1": 933.8, "q3": 2019.6, "volume": 14403, "confidence": 0.692}, "local": {"median": 938.6, "mean": 1189.5, "q1": 590.3, "q3": 1377.8, "volume": 578, "confidence": 0.664}}, "83": {"tous": {"median": 3352.3, "mean": 3779.0, "q1": 2185.2, "q3": 4816.7, "volume": 126979, "confidence": 0.686}, "appartement": {"median": 3529.4, "mean": 3905.7, "q1": 2411.5, "q3": 4991.4, "volume": 73026, "confidence": 0.708}, "maison": {"median": 3214.3, "mean": 3710.1, "q1": 1940.3, "q3": 4675.9, "volume": 49080, "confidence": 0.66}, "local": {"median": 2060.6, "mean": 2575.2, "q1": 1304.3, "q3": 3148.1, "volume": 4873, "confidence": 0.642}}, "84": {"tous": {"median": 2340.4, "mean": 2515.8, "q1": 1644.1, "q3": 3088.2, "volume": 41739, "confidence": 0.753}, "appartement": {"median": 2093.2, "mean": 2197.7, "q1": 1481.3, "q3": 2750.0, "volume": 14836, "confidence": 0.758}, "maison": {"median": 2551.3, "mean": 2760.6, "q1": 1857.1, "q3": 3319.0, "volume": 25053, "confidence": 0.771}, "local": {"median": 1360.6, "mean": 1751.8, "q1": 830.2, "q3": 2193.2, "volume": 1850, "confidence": 0.6}}, "85": {"tous": {"median": 2200.0, "mean": 2456.4, "q1": 1372.5, "q3": 3159.1, "volume": 66695, "confidence": 0.675}, "appartement": {"median": 3030.3, "mean": 3325.6, "q1": 2200.0, "q3": 4170.6, "volume": 11325, "confidence": 0.74}, "maison": {"median": 2080.6, "mean": 2306.9, "q1": 1295.3, "q3": 2945.2, "volume": 52440, "confidence": 0.683}, "local": {"median": 1333.1, "mean": 1773.0, "q1": 703.7, "q3": 2252.5, "volume": 2930, "confidence": 0.6}}, "86": {"tous": {"median": 1309.5, "mean": 1447.7, "q1": 806.5, "q3": 1930.6, "volume": 33289, "confidence": 0.657}, "appartement": {"median": 1963.4, "mean": 1992.8, "q1": 1472.0, "q3": 2496.3, "volume": 6618, "confidence": 0.791}, "maison": {"median": 1164.3, "mean": 1296.7, "q1": 750.0, "q3": 1750.0, "volume": 25387, "confidence": 0.656}, "local": {"median": 964.9, "mean": 1623.3, "q1": 509.5, "q3": 1642.9, "volume": 1284, "confidence": 0.6}}, "87": {"tous": {"median": 1250.0, "mean": 1331.7, "q1": 801.4, "q3": 1759.6, "volume": 28742, "confidence": 0.693}, "appartement": {"median": 1516.1, "mean": 1546.7, "q1": 1182.1, "q3": 1869.6, "volume": 8255, "confidence": 0.819}, "maison": {"median": 1088.5, "mean": 1250.9, "q1": 705.7, "q3": 1687.5, "volume": 19529, "confidence": 0.639}, "local": {"median": 933.3, "mean": 1126.0, "q1": 529.2, "q3": 1503.0, "volume": 958, "confidence": 0.6}}, "88": {"tous": {"median": 1071.0, "mean": 1240.1, "q1": 689.7, "q3": 1580.5, "volume": 25803, "confidence": 0.667}, "appartement": {"median": 1094.8, "mean": 1286.6, "q1": 743.4, "q3": 1568.3, "volume": 8141, "confidence": 0.699}, "maison": {"median": 1075.7, "mean": 1231.5, "q1": 687.5, "q3": 1603.3, "volume": 16612, "confidence": 0.659}, "local": {"median": 750.0, "mean": 1015.5, "q1": 444.4, "q3": 1201.9, "volume": 1050, "confidence": 0.6}}, "89": {"tous": {"median": 1197.9, "mean": 1291.7, "q1": 783.1, "q3": 1671.2, "volume": 26732, "confidence": 0.703}, "appartement": {"median": 1354.2, "mean": 1378.0, "q1": 994.7, "q3": 1723.9, "volume": 4285, "confidence": 0.785}, "maison": {"median": 1172.2, "mean": 1271.4, "q1": 763.6, "q3": 1665.4, "volume": 21553, "confidence": 0.692}, "local": {"median": 849.5, "mean": 1365.5, "q1": 506.8, "q3": 1448.3, "volume": 894, "confidence": 0.6}}, "90": {"tous": {"median": 1333.3, "mean": 1399.7, "q1": 962.3, "q3": 1776.2, "volume": 8960, "confidence": 0.756}, "appartement": {"median": 1214.3, "mean": 1263.9, "q1": 913.6, "q3": 1554.1, "volume": 4715, "confidence": 0.789}, "maison": {"median": 1602.7, "mean": 1605.5, "q1": 1114.8, "q3": 2026.7, "volume": 3850, "confidence": 0.772}, "local": {"median": 824.4, "mean": 1014.8, "q1": 500.0, "q3": 1272.7, "volume": 395, "confidence": 0.625}}, "91": {"tous": {"median": 3165.6, "mean": 3282.5, "q1": 2427.2, "q3": 3921.6, "volume": 77368, "confidence": 0.811}, "appartement": {"median": 2983.3, "mean": 3097.1, "q1": 2347.2, "q3": 3712.1, "volume": 36975, "confidence": 0.817}, "maison": {"median": 3378.4, "mean": 3503.6, "q1": 2674.4, "q3": 4147.5, "volume": 37594, "confidence": 0.826}, "local": {"median": 2027.0, "mean": 2760.2, "q1": 1205.7, "q3": 3333.3, "volume": 2799, "confidence": 0.6}}, "92": {"tous": {"median": 6862.9, "mean": 7100.3, "q1": 5420.9, "q3": 8500.0, "volume": 97813, "confidence": 0.821}, "appartement": {"median": 6823.3, "mean": 7038.5, "q1": 5448.7, "q3": 8375.0, "volume": 82181, "confidence": 0.828}, "maison": {"median": 7822.6, "mean": 8225.7, "q1": 6134.5, "q3": 9820.0, "volume": 11649, "confidence": 0.812}, "local": {"median": 4756.1, "mean": 5083.4, "q1": 2744.2, "q3": 6625.0, "volume": 3983, "confidence": 0.674}}, "93": {"tous": {"median": 3913.0, "mean": 4302.8, "q1": 2947.8, "q3": 5294.1, "volume": 70849, "confidence": 0.76}, "appartement": {"median": 3911.3, "mean": 4287.9, "q1": 2892.9, "q3": 5409.1, "volume": 46907, "confidence": 0.743}, "maison": {"median": 4000.0, "mean": 4430.4, "q1": 3181.8, "q3": 5147.1, "volume": 21420, "confidence": 0.803}, "local": {"median": 2717.4, "mean": 3496.7, "q1": 1625.0, "q3": 4359.0, "volume": 2522, "confidence": 0.6}}, "94": {"tous": {"median": 4900.0, "mean": 5305.3, "q1": 3653.8, "q3": 6466.7, "volume": 76197, "confidence": 0.77}, "appartement": {"median": 4880.9, "mean": 5241.3, "q1": 3658.5, "q3": 6363.6, "volume": 55606, "confidence": 0.778}, "maison": {"median": 5125.0, "mean": 5703.1, "q1": 3854.2, "q3": 7000.0, "volume": 17890, "confidence": 0.754}, "local": {"median": 3378.4, "mean": 3987.9, "q1": 2000.0, "q3": 5294.1, "volume": 2701, "confidence": 0.61}}, "95": {"tous": {"median": 3400.0, "mean": 3530.8, "q1": 2687.4, "q3": 4184.6, "volume": 68727, "confidence": 0.824}, "appartement": {"median": 3226.4, "mean": 3324.0, "q1": 2600.0, "q3": 3942.3, "volume": 33240, "confidence": 0.834}, "maison": {"median": 3611.1, "mean": 3782.1, "q1": 2905.6, "q3": 4447.8, "volume": 33220, "confidence": 0.829}, "local": {"median": 2101.6, "mean": 2881.0, "q1": 1174.7, "q3": 3500.0, "volume": 2267, "confidence": 0.6}}, "971": {"tous": {"median": 2826.1, "mean": 3033.2, "q1": 1743.6, "q3": 3912.1, "volume": 10721, "confidence": 0.693}, "appartement": {"median": 2951.7, "mean": 3120.8, "q1": 2000.0, "q3": 3960.4, "volume": 4988, "confidence": 0.734}, "maison": {"median": 2771.3, "mean": 3011.1, "q1": 1602.2, "q3": 3921.1, "volume": 4326, "confidence": 0.665}, "local": {"median": 2470.0, "mean": 2791.0, "q1": 1395.3, "q3": 3691.3, "volume": 1407, "confidence": 0.628}}, "972": {"tous": {"median": 2500.0, "mean": 2682.8, "q1": 1600.0, "q3": 3460.0, "volume": 9912, "confidence": 0.702}, "appartement": {"median": 2666.7, "mean": 2846.0, "q1": 1966.3, "q3": 3476.9, "volume": 4813, "confidence": 0.773}, "maison": {"median": 2348.7, "mean": 2604.3, "q1": 1322.4, "q3": 3584.4, "volume": 4142, "confidence": 0.615}, "local": {"median": 1925.9, "mean": 2201.6, "q1": 1198.0, "q3": 2872.4, "volume": 957, "confidence": 0.652}}, "973": {"tous": {"median": 2469.1, "mean": 2500.8, "q1": 1760.0, "q3": 3078.9, "volume": 4002, "confidence": 0.786}, "appartement": {"median": 2576.8, "mean": 2603.9, "q1": 2112.7, "q3": 3111.1, "volume": 1586, "confidence": 0.845}, "maison": {"median": 2401.3, "mean": 2478.9, "q1": 1619.8, "q3": 3088.2, "volume": 2260, "confidence": 0.755}, "local": {"median": 1382.9, "mean": 1769.8, "q1": 882.4, "q3": 2036.4, "volume": 156, "confidence": 0.666}}, "974": {"tous": {"median": 2492.1, "mean": 2800.9, "q1": 1833.3, "q3": 3402.8, "volume": 28084, "confidence": 0.748}, "appartement": {"median": 2473.7, "mean": 2781.1, "q1": 1966.7, "q3": 3214.3, "volume": 12306, "confidence": 0.798}, "maison": {"median": 2521.7, "mean": 2829.5, "q1": 1666.7, "q3": 3571.4, "volume": 14873, "confidence": 0.698}, "local": {"median": 2162.2, "mean": 2600.0, "q1": 1386.4, "q3": 3164.6, "volume": 905, "confidence": 0.671}}}

data/aggregated/prices_region.json

@@ -0,0 +1 @@
+{"01": {"tous": {"median": 2826.1, "mean": 3033.2, "q1": 1743.6, "q3": 3912.1, "volume": 10721, "confidence": 0.693}, "appartement": {"median": 2951.7, "mean": 3120.8, "q1": 2000.0, "q3": 3960.4, "volume": 4988, "confidence": 0.734}, "maison": {"median": 2771.3, "mean": 3011.1, "q1": 1602.2, "q3": 3921.1, "volume": 4326, "confidence": 0.665}, "local": {"median": 2470.0, "mean": 2791.0, "q1": 1395.3, "q3": 3691.3, "volume": 1407, "confidence": 0.628}}, "02": {"tous": {"median": 2500.0, "mean": 2682.8, "q1": 1600.0, "q3": 3460.0, "volume": 9912, "confidence": 0.702}, "appartement": {"median": 2666.7, "mean": 2846.0, "q1": 1966.3, "q3": 3476.9, "volume": 4813, "confidence": 0.773}, "maison": {"median": 2348.7, "mean": 2604.3, "q1": 1322.4, "q3": 3584.4, "volume": 4142, "confidence": 0.615}, "local": {"median": 1925.9, "mean": 2201.6, "q1": 1198.0, "q3": 2872.4, "volume": 957, "confidence": 0.652}}, "03": {"tous": {"median": 2469.1, "mean": 2500.8, "q1": 1760.0, "q3": 3078.9, "volume": 4002, "confidence": 0.786}, "appartement": {"median": 2576.8, "mean": 2603.9, "q1": 2112.7, "q3": 3111.1, "volume": 1586, "confidence": 0.845}, "maison": {"median": 2401.3, "mean": 2478.9, "q1": 1619.8, "q3": 3088.2, "volume": 2260, "confidence": 0.755}, "local": {"median": 1382.9, "mean": 1769.8, "q1": 882.4, "q3": 2036.4, "volume": 156, "confidence": 0.666}}, "04": {"tous": {"median": 2492.1, "mean": 2800.9, "q1": 1833.3, "q3": 3402.8, "volume": 28084, "confidence": 0.748}, "appartement": {"median": 2473.7, "mean": 2781.1, "q1": 1966.7, "q3": 3214.3, "volume": 12306, "confidence": 0.798}, "maison": {"median": 2521.7, "mean": 2829.5, "q1": 1666.7, "q3": 3571.4, "volume": 14873, "confidence": 0.698}, "local": {"median": 2162.2, "mean": 2600.0, "q1": 1386.4, "q3": 3164.6, "volume": 905, "confidence": 0.671}}, "11": {"tous": {"median": 4545.5, "mean": 5730.9, "q1": 3077.3, "q3": 7896.2, "volume": 729713, "confidence": 0.6}, "appartement": {"median": 5606.1, "mean": 6527.2, "q1": 3481.5, "q3": 9145.5, "volume": 481565, "confidence": 0.6}, "maison": {"median": 3529.4, "mean": 4019.9, "q1": 2593.4, "q3": 4755.6, "volume": 215650, "confidence": 0.755}, "local": {"median": 3975.0, "mean": 5285.5, "q1": 1916.7, "q3": 7640.4, "volume": 32498, "confidence": 0.6}}, "24": {"tous": {"median": 1489.4, "mean": 1667.3, "q1": 916.7, "q3": 2209.6, "volume": 192214, "confidence": 0.653}, "appartement": {"median": 2045.5, "mean": 2126.8, "q1": 1428.6, "q3": 2750.0, "volume": 43364, "confidence": 0.742}, "maison": {"median": 1344.8, "mean": 1533.8, "q1": 845.1, "q3": 2015.4, "volume": 142096, "confidence": 0.652}, "local": {"median": 1000.0, "mean": 1524.8, "q1": 545.5, "q3": 1829.8, "volume": 6754, "confidence": 0.6}}, "27": {"tous": {"median": 1392.9, "mean": 1578.7, "q1": 869.0, "q3": 2099.0, "volume": 208712, "confidence": 0.647}, "appartement": {"median": 1659.7, "mean": 1797.3, "q1": 1062.5, "q3": 2411.8, "volume": 71719, "confidence": 0.675}, "maison": {"median": 1290.3, "mean": 1471.6, "q1": 807.7, "q3": 1935.5, "volume": 128640, "confidence": 0.65}, "local": {"median": 933.3, "mean": 1351.1, "q1": 515.4, "q3": 1637.9, "volume": 8353, "confidence": 0.6}}, "28": {"tous": {"median": 1768.4, "mean": 1984.7, "q1": 1094.1, "q3": 2543.3, "volume": 258639, "confidence": 0.672}, "appartement": {"median": 2305.6, "mean": 2561.8, "q1": 1612.9, "q3": 3192.4, "volume": 72471, "confidence": 0.726}, "maison": {"median": 1574.5, "mean": 1761.8, "q1": 982.1, "q3": 2288.6, "volume": 177132, "confidence": 0.668}, "local": {"median": 1215.5, "mean": 1726.8, "q1": 654.2, "q3": 2217.0, "volume": 9036, "confidence": 0.6}}, "32": {"tous": {"median": 1735.8, "mean": 1984.7, "q1": 1145.8, "q3": 2500.0, "volume": 382078, "confidence": 0.688}, "appartement": {"median": 2333.3, "mean": 2634.0, "q1": 1557.4, "q3": 3362.1, "volume": 83415, "confidence": 0.691}, "maison": {"median": 1625.0, "mean": 1802.6, "q1": 1082.6, "q3": 2285.7, "volume": 285981, "confidence": 0.704}, "local": {"median": 1190.5, "mean": 1820.0, "q1": 642.2, "q3": 2210.5, "volume": 12682, "confidence": 0.6}}, "44": {"tous": {"median": 1417.3, "mean": 1564.0, "q1": 884.6, "q3": 2071.4, "volume": 176812, "confidence": 0.665}, "appartement": {"median": 1631.1, "mean": 1763.5, "q1": 1101.7, "q3": 2286.3, "volume": 59150, "confidence": 0.709}, "maison": {"median": 1324.3, "mean": 1468.1, "q1": 809.3, "q3": 1961.9, "volume": 110836, "confidence": 0.652}, "local": {"median": 966.9, "mean": 1392.8, "q1": 550.8, "q3": 1649.5, "volume": 6826, "confidence": 0.6}}, "52": {"tous": {"median": 2160.0, "mean": 2435.8, "q1": 1319.6, "q3": 3219.2, "volume": 301827, "confidence": 0.648}, "appartement": {"median": 3005.7, "mean": 3187.4, "q1": 2125.0, "q3": 3958.3, "volume": 79110, "confidence": 0.756}, "maison": {"median": 1917.1, "mean": 2187.7, "q1": 1173.5, "q3": 2841.5, "volume": 210441, "confidence": 0.652}, "local": {"median": 1347.0, "mean": 1845.0, "q1": 699.3, "q3": 2381.0, "volume": 12276, "confidence": 0.6}}, "53": {"tous": {"median": 2033.0, "mean": 2283.4, "q1": 1261.7, "q3": 2948.7, "volume": 280003, "confidence": 0.668}, "appartement": {"median": 2614.4, "mean": 2856.4, "q1": 1868.3, "q3": 3613.6, "volume": 80626, "confidence": 0.733}, "maison": {"median": 1829.3, "mean": 2076.7, "q1": 1111.1, "q3": 2666.7, "volume": 187994, "confidence": 0.66}, "local": {"median": 1255.8, "mean": 1637.6, "q1": 712.5, "q3": 2142.9, "volume": 11383, "confidence": 0.6}}, "75": {"tous": {"median": 1853.8, "mean": 2415.6, "q1": 1072.1, "q3": 3248.8, "volume": 505650, "confidence": 0.6}, "appartement": {"median": 2859.1, "mean": 3258.9, "q1": 1756.7, "q3": 4290.3, "volume": 139810, "confidence": 0.646}, "maison": {"median": 1586.2, "mean": 2099.2, "q1": 936.6, "q3": 2652.9, "volume": 347220, "confidence": 0.6}, "local": {"median": 1343.4, "mean": 1984.4, "q1": 676.7, "q3": 2547.2, "volume": 18620, "confidence": 0.6}}, "76": {"tous": {"median": 2125.0, "mean": 2356.1, "q1": 1300.0, "q3": 3095.2, "volume": 523730, "confidence": 0.662}, "appartement": {"median": 2520.9, "mean": 2721.5, "q1": 1629.6, "q3": 3531.2, "volume": 213001, "confidence": 0.698}, "maison": {"median": 1900.0, "mean": 2129.7, "q1": 1166.7, "q3": 2804.6, "volume": 290513, "confidence": 0.655}, "local": {"median": 1353.2, "mean": 1760.7, "q1": 746.1, "q3": 2254.5, "volume": 20216, "confidence": 0.6}}, "84": {"tous": {"median": 2391.3, "mean": 2785.3, "q1": 1434.8, "q3": 3673.5, "volume": 606247, "confidence": 0.626}, "appartement": {"median": 2804.3, "mean": 3162.9, "q1": 1756.1, "q3": 4148.9, "volume": 310845, "confidence": 0.659}, "maison": {"median": 2086.6, "mean": 2429.4, "q1": 1238.7, "q3": 3126.3, "volume": 264552, "confidence": 0.638}, "local": {"median": 1414.4, "mean": 2033.1, "q1": 789.1, "q3": 2527.8, "volume": 30850, "confidence": 0.6}}, "93": {"tous": {"median": 3381.3, "mean": 3761.0, "q1": 2263.1, "q3": 4736.8, "volume": 493722, "confidence": 0.707}, "appartement": {"median": 3546.9, "mean": 3872.9, "q1": 2419.4, "q3": 4862.4, "volume": 314937, "confidence": 0.724}, "maison": {"median": 3220.2, "mean": 3693.3, "q1": 2140.2, "q3": 4588.2, "volume": 155197, "confidence": 0.696}, "local": {"median": 2155.5, "mean": 2711.4, "q1": 1305.6, "q3": 3366.7, "volume": 23588, "confidence": 0.618}}, "94": {"tous": {"median": 3166.7, "mean": 3486.4, "q1": 2169.2, "q3": 4375.0, "volume": 23090, "confidence": 0.721}, "appartement": {"median": 3200.1, "mean": 3410.3, "q1": 2333.3, "q3": 4317.2, "volume": 14176, "confidence": 0.752}, "maison": {"median": 3274.6, "mean": 3818.7, "q1": 2037.0, "q3": 4727.3, "volume": 7499, "confidence": 0.671}, "local": {"median": 2146.3, "mean": 2487.5, "q1": 1280.0, "q3": 3250.0, "volume": 1415, "confidence": 0.633}}}

data/aggregated/top_cities.json

@@ -0,0 +1,281 @@
+{
+  "Bordeaux": {
+    "code": "33063",
+    "appartement": {
+      "median": 4439.2,
+      "mean": 4566.7,
+      "q1": 3653.8,
+      "q3": 5333.3,
+      "volume": 18352
+    },
+    "local": {
+      "median": 3487.2,
+      "mean": 4017.4,
+      "q1": 2343.5,
+      "q3": 4945.1,
+      "volume": 1218
+    },
+    "maison": {
+      "median": 5147.0,
+      "mean": 5317.0,
+      "q1": 4210.5,
+      "q3": 6153.8,
+      "volume": 6018
+    },
+    "tous": {
+      "median": 4566.4,
+      "mean": 4717.0,
+      "q1": 3697.8,
+      "q3": 5543.5,
+      "volume": 25588
+    }
+  },
+  "Lille": {
+    "code": "59350",
+    "appartement": {
+      "median": 3673.5,
+      "mean": 3774.2,
+      "q1": 2897.4,
+      "q3": 4517.9,
+      "volume": 14741
+    },
+    "local": {
+      "median": 2542.9,
+      "mean": 3167.6,
+      "q1": 1563.9,
+      "q3": 3812.5,
+      "volume": 1064
+    },
+    "maison": {
+      "median": 2785.8,
+      "mean": 2930.7,
+      "q1": 2172.4,
+      "q3": 3500.0,
+      "volume": 4911
+    },
+    "tous": {
+      "median": 3405.1,
+      "mean": 3543.1,
+      "q1": 2600.0,
+      "q3": 4312.5,
+      "volume": 20716
+    }
+  },
+  "Lyon": {
+    "code": "69123",
+    "appartement": {
+      "median": 4826.2,
+      "mean": 4830.9,
+      "q1": 3875.0,
+      "q3": 5711.9,
+      "volume": 35236
+    },
+    "local": {
+      "median": 3311.4,
+      "mean": 3753.1,
+      "q1": 2173.9,
+      "q3": 4703.4,
+      "volume": 2734
+    },
+    "maison": {
+      "median": 5990.0,
+      "mean": 6463.1,
+      "q1": 4662.2,
+      "q3": 7609.6,
+      "volume": 949
+    },
+    "tous": {
+      "median": 4777.8,
+      "mean": 4795.0,
+      "q1": 3756.5,
+      "q3": 5710.7,
+      "volume": 38919
+    }
+  },
+  "Marseille": {
+    "code": "13055",
+    "appartement": {
+      "median": 2935.5,
+      "mean": 3110.5,
+      "q1": 2089.6,
+      "q3": 3909.1,
+      "volume": 59629
+    },
+    "local": {
+      "median": 2062.5,
+      "mean": 2473.1,
+      "q1": 1279.1,
+      "q3": 3157.9,
+      "volume": 3731
+    },
+    "maison": {
+      "median": 4505.3,
+      "mean": 4990.0,
+      "q1": 3283.3,
+      "q3": 6014.7,
+      "volume": 7871
+    },
+    "tous": {
+      "median": 3023.8,
+      "mean": 3284.8,
+      "q1": 2111.1,
+      "q3": 4083.3,
+      "volume": 71231
+    }
+  },
+  "Montpellier": {
+    "code": "34172",
+    "appartement": {
+      "median": 3261.3,
+      "mean": 3280.2,
+      "q1": 2522.7,
+      "q3": 4000.0,
+      "volume": 22452
+    },
+    "local": {
+      "median": 2564.1,
+      "mean": 2766.4,
+      "q1": 1639.3,
+      "q3": 3466.8,
+      "volume": 1585
+    },
+    "maison": {
+      "median": 3819.5,
+      "mean": 4073.3,
+      "q1": 3070.7,
+      "q3": 4806.0,
+      "volume": 2510
+    },
+    "tous": {
+      "median": 3279.5,
+      "mean": 3324.5,
+      "q1": 2508.0,
+      "q3": 4047.6,
+      "volume": 26547
+    }
+  },
+  "Nantes": {
+    "code": "44109",
+    "appartement": {
+      "median": 3690.5,
+      "mean": 3724.2,
+      "q1": 3000.0,
+      "q3": 4393.4,
+      "volume": 21661
+    },
+    "local": {
+      "median": 2911.7,
+      "mean": 3252.4,
+      "q1": 1845.2,
+      "q3": 4227.3,
+      "volume": 1334
+    },
+    "maison": {
+      "median": 4433.6,
+      "mean": 4675.4,
+      "q1": 3636.4,
+      "q3": 5500.0,
+      "volume": 5544
+    },
+    "tous": {
+      "median": 3793.1,
+      "mean": 3886.9,
+      "q1": 3048.8,
+      "q3": 4575.8,
+      "volume": 28539
+    }
+  },
+  "Nice": {
+    "code": "06088",
+    "appartement": {
+      "median": 4444.4,
+      "mean": 4717.2,
+      "q1": 3478.3,
+      "q3": 5600.0,
+      "volume": 39206
+    },
+    "local": {
+      "median": 3214.3,
+      "mean": 3623.2,
+      "q1": 2054.8,
+      "q3": 4722.2,
+      "volume": 2412
+    },
+    "maison": {
+      "median": 5312.5,
+      "mean": 5688.8,
+      "q1": 3689.0,
+      "q3": 6964.6,
+      "volume": 1565
+    },
+    "tous": {
+      "median": 4416.7,
+      "mean": 4691.3,
+      "q1": 3409.1,
+      "q3": 5616.4,
+      "volume": 43183
+    }
+  },
+  "Paris": {
+    "code": "75056",
+    "appartement": {
+      "median": 10362.8,
+      "mean": 10477.6,
+      "q1": 8709.5,
+      "q3": 12106.8,
+      "volume": 147088
+    },
+    "local": {
+      "median": 8255.6,
+      "mean": 8564.9,
+      "q1": 5411.0,
+      "q3": 11215.0,
+      "volume": 11620
+    },
+    "maison": {
+      "median": 13457.8,
+      "mean": 13777.4,
+      "q1": 10400.0,
+      "q3": 17260.3,
+      "volume": 652
+    },
+    "tous": {
+      "median": 10284.5,
+      "mean": 10351.6,
+      "q1": 8541.6,
+      "q3": 12083.3,
+      "volume": 159360
+    }
+  },
+  "Toulouse": {
+    "code": "31555",
+    "appartement": {
+      "median": 3230.8,
+      "mean": 3420.9,
+      "q1": 2552.2,
+      "q3": 4124.3,
+      "volume": 35114
+    },
+    "local": {
+      "median": 2500.0,
+      "mean": 2990.2,
+      "q1": 1580.5,
+      "q3": 3766.8,
+      "volume": 1845
+    },
+    "maison": {
+      "median": 3750.0,
+      "mean": 4100.6,
+      "q1": 2881.3,
+      "q3": 4969.7,
+      "volume": 6119
+    },
+    "tous": {
+      "median": 3272.7,
+      "mean": 3499.0,
+      "q1": 2557.3,
+      "q3": 4224.7,
+      "volume": 43078
+    }
+  }
+}

main.py

@@ -0,0 +1,6 @@
+def main():
+    print("Hello from realadvisor!")
+
+
+if __name__ == "__main__":
+    main()

ml_challenge.txt

@@ -0,0 +1,36 @@
+MLE Challenge
+Context
+You are analysing residential property prices in France, you have at your disposal the public recent transactions on (https://www.data.gouv.fr/datasets/demandes-de-valeurs-foncieres/) to estimate the current market price as price per squared meter (€/m²)
+Objective
+You need to generate an interactive map visualization for the aggregated price data by : 
+Country
+Region
+Departament
+Neighborhood
+Postcode
+Building plots
+
+You need to procure yourself with open data from the french government for such geometries and take care of cleaning the data to make accurate aggregates. 
+You are asked to make the best estimation of the market price taking into account, transaction price volatility, transaction volume, data freshness and consistency. You can estimate a number for the price or an interval.
+You can use this one as a reference https://explore.data.gouv.fr/fr/immobilier
+
+You need to render an interactive map that shows the price aggregate with colors, as the level zooms, it needs to transition between aggregation levels. (Link to hosted app, optional but preferred)
+If the volume of data is too important for the browser to support all you can subset it, but a solution for this will be appreciated. 
+Produce a list of market price per square meter by property type for the top 10 biggest cities.
+Submit your processing code. (Link to github repo) 
+What you will be evaluated on 
+Is the colored map loading ?
+Is the map usable and not laggy ?
+Is the map refreshing the aggregation level on zoom ?
+Are all 6 aggregation levels present ?
+Country
+Region
+Departament
+Neighborhood
+Postcode
+Building plots
+Are the price estimates plausible ?
+Is the data complete or was it subset ?
+The processing code is clean, clear and reusable
+The architecture is robust and logical
+App is hosted and functional

pyproject.toml

@@ -0,0 +1,17 @@
+[project]
+name = "realadvisor-mle-challenge"
+version = "0.1.0"
+description = "French property price analysis and interactive map visualization"
+readme = "README.md"
+requires-python = ">=3.11"
+dependencies = [
+    "polars>=1.0.0",
+    "requests>=2.31.0",
+    "tqdm>=4.66.0",
+    "jupyter>=1.0.0",
+    "matplotlib>=3.8.0",
+    "seaborn>=0.13.0",
+    "geopandas>=0.14.0",
+    "folium>=0.15.0",
+    "lab>=8.8",
+]

src/__init__.py

@@ -0,0 +1 @@
+"""RealAdvisor MLE Challenge - French property price analysis pipeline."""

src/aggregator.py

@@ -0,0 +1,194 @@
+"""
+Aggregate property prices at different geographic levels.
+
+Responsibility: Given cleaned transaction data, compute summary statistics
+(median, volume, IQR, confidence) grouped by any geographic column.
+"""
+
+import json
+import logging
+from pathlib import Path
+
+import polars as pl
+
+from src.config import (
+    AGGREGATED_DIR,
+    AGGREGATION_LEVELS,
+    DEPT_TO_REGION,
+    NO_DVF_DEPARTMENTS,
+    REGION_NAMES,
+    TYPE_LOCAL_SHORT,
+)
+
+logger = logging.getLogger(__name__)
+
+# Map each aggregation level to the column used for grouping
+LEVEL_TO_COLUMN: dict[str, str] = {
+    "country": "_country",          # synthetic constant column
+    "region": "code_region",
+    "department": "code_departement",
+    "commune": "code_commune",
+    "postcode": "code_postal",
+    "section": "code_section",
+}
+
+
+def aggregate_level(
+    lf: pl.LazyFrame,
+    group_col: str,
+    *,
+    property_type: str | None = None,
+) -> pl.DataFrame:
+    """
+    Compute price statistics for one geographic level and property type.
+
+    Args:
+        lf:            Cleaned transaction LazyFrame.
+        group_col:     Column to group by (e.g. "code_departement").
+        property_type: Filter to this type_local value, or None for all.
+
+    Returns:
+        DataFrame with columns: code, median_price_m2, mean_price_m2,
+        q1, q3, volume, std_dev.
+    """
+    filtered = lf
+    if property_type:
+        filtered = filtered.filter(pl.col("type_local") == property_type)
+
+    # For country level, add a constant column
+    if group_col == "_country":
+        filtered = filtered.with_columns(pl.lit("FR").alias("_country"))
+
+    return (
+        filtered.group_by(group_col)
+        .agg(
+            pl.col("prix_m2").median().alias("median_price_m2"),
+            pl.col("prix_m2").mean().alias("mean_price_m2"),
+            pl.col("prix_m2").quantile(0.25).alias("q1"),
+            pl.col("prix_m2").quantile(0.75).alias("q3"),
+            pl.col("prix_m2").std().alias("std_dev"),
+            pl.col("prix_m2").count().alias("volume"),
+            pl.col("year").cast(pl.Int32, strict=False).max().alias("latest_year"),
+        )
+        .rename({group_col: "code"})
+        .with_columns(
+            # IQR = Q3 - Q1
+            (pl.col("q3") - pl.col("q1")).alias("iqr"),
+        )
+        .sort("code")
+        .collect()
+    )
+
+
+def compute_confidence(df: pl.DataFrame) -> pl.DataFrame:
+    """
+    Add a confidence score (0-1) based on volume and volatility.
+
+    Higher volume + lower relative IQR = higher confidence.
+    """
+    return df.with_columns(
+        (
+            # Volume component: log-scaled, saturates around 100 transactions
+            pl.col("volume").cast(pl.Float64).log1p() / pl.lit(100.0).log1p()
+        ).clip(0.0, 1.0).alias("conf_volume"),
+        (
+            # Stability component: 1 - (IQR / median), clamped
+            pl.when(pl.col("median_price_m2") > 0)
+            .then(
+                (1.0 - pl.col("iqr") / pl.col("median_price_m2")).clip(0.0, 1.0)
+            )
+            .otherwise(0.0)
+        ).alias("conf_stability"),
+    ).with_columns(
+        # Overall confidence = weighted average
+        (pl.col("conf_volume") * 0.6 + pl.col("conf_stability") * 0.4)
+        .round(3)
+        .alias("confidence")
+    )
+
+
+def aggregate_all_types(
+    lf: pl.LazyFrame,
+    group_col: str,
+) -> pl.DataFrame:
+    """
+    Aggregate for all property types + combined "tous" for a given level.
+
+    Returns:
+        DataFrame with columns: code, type, median_price_m2, volume, confidence, ...
+    """
+    results = []
+
+    # All types combined
+    df_all = aggregate_level(lf, group_col)
+    df_all = compute_confidence(df_all)
+    df_all = df_all.with_columns(pl.lit("tous").alias("type"))
+    results.append(df_all)
+
+    # Per property type
+    for full_name, short_name in TYPE_LOCAL_SHORT.items():
+        df_type = aggregate_level(lf, group_col, property_type=full_name)
+        df_type = compute_confidence(df_type)
+        df_type = df_type.with_columns(pl.lit(short_name).alias("type"))
+        results.append(df_type)
+
+    return pl.concat(results, how="vertical_relaxed")
+
+
+def aggregate_all_levels(lf: pl.LazyFrame) -> dict[str, pl.DataFrame]:
+    """
+    Run aggregation for all 6 geographic levels.
+
+    Args:
+        lf: Cleaned LazyFrame.
+
+    Returns:
+        Dict mapping level name → aggregated DataFrame.
+    """
+    results = {}
+    for level in AGGREGATION_LEVELS:
+        col = LEVEL_TO_COLUMN[level]
+        logger.info("Aggregating level: %s (group by %s)", level, col)
+        results[level] = aggregate_all_types(lf, col)
+        logger.info(
+            "  → %d rows (%d unique codes)",
+            len(results[level]),
+            results[level]["code"].n_unique(),
+        )
+    return results
+
+
+def export_json(aggregated: dict[str, pl.DataFrame], output_dir: Path | None = None) -> None:
+    """
+    Export aggregated DataFrames to JSON files for the frontend.
+
+    Each level produces a JSON file structured as:
+    { "code1": { "tous": {...}, "appartement": {...}, ... }, ... }
+
+    Args:
+        aggregated: Dict from aggregate_all_levels().
+        output_dir: Directory to write JSONs. Defaults to config.AGGREGATED_DIR.
+    """
+    output_dir = output_dir or AGGREGATED_DIR
+    output_dir.mkdir(parents=True, exist_ok=True)
+
+    for level, df in aggregated.items():
+        data: dict = {}
+        for row in df.iter_rows(named=True):
+            code = str(row["code"])
+            ptype = row["type"]
+            if code not in data:
+                data[code] = {}
+            data[code][ptype] = {
+                "median": round(row["median_price_m2"] or 0, 1),
+                "mean": round(row["mean_price_m2"] or 0, 1),
+                "q1": round(row["q1"] or 0, 1),
+                "q3": round(row["q3"] or 0, 1),
+                "volume": row["volume"],
+                "confidence": row["confidence"],
+            }
+
+        path = output_dir / f"prices_{level}.json"
+        with open(path, "w") as f:
+            json.dump(data, f, ensure_ascii=False)
+        logger.info("Exported: %s (%d entries)", path.name, len(data))

src/cleaner.py

@@ -0,0 +1,245 @@
+"""
+Clean and filter raw DVF transaction data.
+
+Responsibility: Load raw CSVs, apply quality filters, deduplicate
+multi-row mutations, and produce a clean dataset ready for price calculation.
+"""
+
+import logging
+from pathlib import Path
+
+import polars as pl
+
+from src.config import (
+    ARRONDISSEMENT_MAPPING,
+    DVF_COLUMNS,
+    PRICE_M2_MAX,
+    PRICE_M2_MIN,
+    REFERENCE_DATE,
+    SURFACE_MAX,
+    SURFACE_MIN,
+    TEMPORAL_LAMBDA,
+    VALID_NATURE_MUTATION,
+    VALID_TYPE_LOCAL,
+)
+
+logger = logging.getLogger(__name__)
+
+
+def load_raw_csv(path: Path) -> pl.LazyFrame:
+    """
+    Load a single raw DVF CSV into a Polars LazyFrame.
+
+    Args:
+        path: Path to the CSV file (decompressed).
+
+    Returns:
+        LazyFrame with only the columns we need.
+    """
+    # Force string types for geographic codes (Corsica has 2A/2B prefixes)
+    schema_overrides = {
+        "id_mutation": pl.Utf8,
+        "code_postal": pl.Utf8,
+        "code_commune": pl.Utf8,
+        "code_departement": pl.Utf8,
+        "id_parcelle": pl.Utf8,
+        "code_type_local": pl.Utf8,
+        "date_mutation": pl.Utf8,
+        "nom_commune": pl.Utf8,
+        "nature_mutation": pl.Utf8,
+        "type_local": pl.Utf8,
+    }
+    return pl.scan_csv(
+        path,
+        separator=",",
+        infer_schema_length=10_000,
+        null_values=["", "NA", "null"],
+        schema_overrides=schema_overrides,
+    ).select(DVF_COLUMNS)
+
+
+def load_multiple_csvs(paths: list[Path]) -> pl.LazyFrame:
+    """
+    Load and concatenate multiple DVF CSV files.
+
+    Args:
+        paths: List of paths to CSV files.
+
+    Returns:
+        Single LazyFrame with all years combined.
+    """
+    frames = [load_raw_csv(p) for p in paths]
+    return pl.concat(frames, how="vertical_relaxed")
+
+
+def filter_sales(lf: pl.LazyFrame) -> pl.LazyFrame:
+    """
+    Keep only standard property sales.
+
+    Filters:
+        - nature_mutation == "Vente"
+        - type_local in [Appartement, Maison, Local industriel. et commercial.]
+        - valeur_fonciere > 0
+        - surface_reelle_bati > 0
+    """
+    return lf.filter(
+        (pl.col("nature_mutation") == VALID_NATURE_MUTATION)
+        & (pl.col("type_local").is_in(VALID_TYPE_LOCAL))
+        & (pl.col("valeur_fonciere").is_not_null())
+        & (pl.col("valeur_fonciere") > 0)
+        & (pl.col("surface_reelle_bati").is_not_null())
+        & (pl.col("surface_reelle_bati") > 0)
+    )
+
+
+def deduplicate_mutations(lf: pl.LazyFrame) -> pl.LazyFrame:
+    """
+    Handle multi-row mutations in DVF data.
+
+    A single sale (id_mutation) can span multiple rows when multiple lots
+    are included. We:
+      1. Keep only mutations where all rows share the same type_local
+         (mixed-type sales have ambiguous price attribution).
+      2. Group by id_mutation: sum surface, keep price (same across rows).
+
+    Returns:
+        One row per mutation with aggregated surface.
+    """
+    # Tag mutations that have a single property type
+    type_counts = lf.group_by("id_mutation").agg(
+        pl.col("type_local").n_unique().alias("n_types")
+    )
+    single_type = type_counts.filter(pl.col("n_types") == 1).select("id_mutation")
+
+    # Keep only single-type mutations, then aggregate
+    filtered = lf.join(single_type, on="id_mutation", how="inner")
+
+    return filtered.group_by("id_mutation").agg(
+        pl.col("date_mutation").first(),
+        pl.col("nature_mutation").first(),
+        pl.col("valeur_fonciere").first(),
+        pl.col("code_postal").first(),
+        pl.col("code_commune").first(),
+        pl.col("nom_commune").first(),
+        pl.col("code_departement").first(),
+        pl.col("id_parcelle").first(),
+        pl.col("type_local").first(),
+        pl.col("surface_reelle_bati").sum(),
+        pl.col("nombre_pieces_principales").sum(),
+        pl.col("nombre_lots").first(),
+        pl.col("longitude").first(),
+        pl.col("latitude").first(),
+    )
+
+
+def normalize_commune_codes(lf: pl.LazyFrame) -> pl.LazyFrame:
+    """
+    Map arrondissement codes back to parent city codes.
+
+    Paris (75101-75120 → 75056), Lyon (69381-69389 → 69123),
+    Marseille (13201-13216 → 13055).
+    """
+    mapping_expr = pl.col("code_commune").cast(pl.Utf8)
+    for arr_code, city_code in ARRONDISSEMENT_MAPPING.items():
+        mapping_expr = (
+            pl.when(pl.col("code_commune").cast(pl.Utf8) == arr_code)
+            .then(pl.lit(city_code))
+            .otherwise(mapping_expr)
+        )
+
+    return lf.with_columns(
+        mapping_expr.alias("code_commune_city"),
+        # Keep original for section-level work
+        pl.col("code_commune").alias("code_commune_original"),
+    )
+
+
+def remove_outliers(lf: pl.LazyFrame) -> pl.LazyFrame:
+    """
+    Remove transactions with implausible price/m² or surface values.
+
+    Uses absolute thresholds from config — intentionally conservative
+    to avoid removing valid luxury or rural transactions.
+    """
+    return lf.filter(
+        (pl.col("surface_reelle_bati") >= SURFACE_MIN)
+        & (pl.col("surface_reelle_bati") <= SURFACE_MAX)
+        & (pl.col("prix_m2") >= PRICE_M2_MIN)
+        & (pl.col("prix_m2") <= PRICE_M2_MAX)
+    )
+
+
+def add_derived_columns(lf: pl.LazyFrame) -> pl.LazyFrame:
+    """
+    Add computed columns needed downstream.
+
+    Adds:
+        - prix_m2: valeur_fonciere / surface_reelle_bati
+        - code_section: first 10 chars of id_parcelle
+        - code_region: mapped from code_departement
+        - year: extracted from date_mutation
+        - months_since: months between transaction and reference date
+        - temporal_weight: exponential decay weight for time-relevance
+    """
+    from src.config import DEPT_TO_REGION
+
+    # Build region mapping expression
+    region_expr = pl.lit("unknown")
+    for dept, region in DEPT_TO_REGION.items():
+        region_expr = (
+            pl.when(pl.col("code_departement").cast(pl.Utf8) == dept)
+            .then(pl.lit(region))
+            .otherwise(region_expr)
+        )
+
+    ref_date = pl.lit(REFERENCE_DATE).str.to_date("%Y-%m-%d")
+
+    return lf.with_columns(
+        (pl.col("valeur_fonciere") / pl.col("surface_reelle_bati")).alias("prix_m2"),
+        pl.col("id_parcelle").cast(pl.Utf8).str.slice(0, 10).alias("code_section"),
+        region_expr.alias("code_region"),
+        pl.col("date_mutation").cast(pl.Utf8).str.slice(0, 4).alias("year"),
+        # Temporal weighting: months since reference date
+        (
+            (ref_date - pl.col("date_mutation").cast(pl.Utf8).str.to_date("%Y-%m-%d"))
+            .dt.total_days()
+            .cast(pl.Float64)
+            / 30.44  # average days per month
+        ).alias("months_since"),
+    ).with_columns(
+        # Exponential decay weight
+        (pl.lit(TEMPORAL_LAMBDA) ** pl.col("months_since").clip(0.0, None)).alias(
+            "temporal_weight"
+        ),
+    )
+
+
+def clean(paths: list[Path]) -> pl.LazyFrame:
+    """
+    Run the full cleaning pipeline on raw DVF files.
+
+    Args:
+        paths: List of raw CSV file paths.
+
+    Returns:
+        Cleaned LazyFrame ready for aggregation.
+    """
+    logger.info("Loading %d raw files...", len(paths))
+    lf = load_multiple_csvs(paths)
+
+    logger.info("Filtering sales...")
+    lf = filter_sales(lf)
+
+    logger.info("Deduplicating mutations...")
+    lf = deduplicate_mutations(lf)
+
+    logger.info("Adding derived columns...")
+    lf = add_derived_columns(lf)
+
+    logger.info("Removing outliers...")
+    lf = remove_outliers(lf)
+
+    logger.info("Normalizing commune codes...")
+    lf = normalize_commune_codes(lf)
+
+    return lf

src/config.py

@@ -0,0 +1,149 @@
+"""
+Configuration constants for the DVF data pipeline.
+
+Single source of truth for paths, URLs, thresholds, and mappings.
+"""
+
+from pathlib import Path
+
+# ---------------------------------------------------------------------------
+# Paths
+# ---------------------------------------------------------------------------
+ROOT_DIR = Path(__file__).resolve().parent.parent
+DATA_DIR = ROOT_DIR / "data"
+RAW_DIR = DATA_DIR / "raw"
+PROCESSED_DIR = DATA_DIR / "processed"
+AGGREGATED_DIR = DATA_DIR / "aggregated"
+SECTIONS_DIR = AGGREGATED_DIR / "sections"
+
+# ---------------------------------------------------------------------------
+# DVF data source
+# ---------------------------------------------------------------------------
+DVF_BASE_URL = "https://files.data.gouv.fr/geo-dvf/latest/csv"
+DVF_YEARS = list(range(2014, 2026))  # 2014-2025: full dataset per Carlos's feedback
+
+def dvf_url(year: int) -> str:
+    """Return download URL for a given year's national DVF geolocalized CSV."""
+    return f"{DVF_BASE_URL}/{year}/full.csv.gz"
+
+# ---------------------------------------------------------------------------
+# Columns we actually need (saves memory on load)
+# ---------------------------------------------------------------------------
+DVF_COLUMNS = [
+    "id_mutation",
+    "date_mutation",
+    "nature_mutation",
+    "valeur_fonciere",
+    "code_postal",
+    "code_commune",
+    "nom_commune",
+    "code_departement",
+    "id_parcelle",
+    "code_type_local",
+    "type_local",
+    "surface_reelle_bati",
+    "nombre_pieces_principales",
+    "nombre_lots",
+    "longitude",
+    "latitude",
+]
+
+# ---------------------------------------------------------------------------
+# Filtering thresholds
+# ---------------------------------------------------------------------------
+VALID_NATURE_MUTATION = "Vente"
+
+VALID_TYPE_LOCAL = ["Appartement", "Maison"]  # Residential only per Carlos's feedback
+
+TYPE_LOCAL_SHORT = {
+    "Appartement": "appartement",
+    "Maison": "maison",
+}
+
+# ---------------------------------------------------------------------------
+# Temporal weighting parameters
+# ---------------------------------------------------------------------------
+REFERENCE_DATE = "2025-01-01"  # Anchor date for temporal decay
+TEMPORAL_LAMBDA = 0.97         # Monthly decay factor (half-life ~23 months)
+TRIM_FRACTION = 0.20           # Trim 20% from each tail for trimmed mean
+
+# Price per m² bounds for outlier removal
+PRICE_M2_MIN = 200       # €/m² — below this is almost certainly an error
+PRICE_M2_MAX = 25_000    # €/m² — above this is extreme luxury / error
+SURFACE_MIN = 9           # m² — below 9m² is legally not habitable in France
+SURFACE_MAX = 1000        # m² — above this for a single unit is suspect
+
+# ---------------------------------------------------------------------------
+# Department → Region mapping (2016 reform)
+# ---------------------------------------------------------------------------
+DEPT_TO_REGION: dict[str, str] = {}
+_REGION_DEPTS = {
+    "84-Auvergne-Rhône-Alpes": "01,03,07,15,26,38,42,43,63,69,73,74",
+    "27-Bourgogne-Franche-Comté": "21,25,39,58,70,71,89,90",
+    "53-Bretagne": "22,29,35,56",
+    "24-Centre-Val de Loire": "18,28,36,37,41,45",
+    "94-Corse": "2A,2B",
+    "44-Grand Est": "08,10,51,52,54,55,57,67,68,88",
+    "32-Hauts-de-France": "02,59,60,62,80",
+    "11-Île-de-France": "75,77,78,91,92,93,94,95",
+    "28-Normandie": "14,27,50,61,76",
+    "75-Nouvelle-Aquitaine": "16,17,19,23,24,33,40,47,64,79,86,87",
+    "76-Occitanie": "09,11,12,30,31,32,34,46,48,65,66,81,82",
+    "52-Pays de la Loire": "44,49,53,72,85",
+    "93-Provence-Alpes-Côte d'Azur": "04,05,06,13,83,84",
+    "01-Guadeloupe": "971",
+    "02-Martinique": "972",
+    "03-Guyane": "973",
+    "04-La Réunion": "974",
+    "06-Mayotte": "976",
+}
+
+REGION_NAMES: dict[str, str] = {}
+for key, depts_str in _REGION_DEPTS.items():
+    code, name = key.split("-", 1)
+    REGION_NAMES[code] = name
+    for d in depts_str.split(","):
+        DEPT_TO_REGION[d.strip()] = code
+
+# Departments with no DVF data (Alsace-Moselle + Mayotte)
+NO_DVF_DEPARTMENTS = {"57", "67", "68", "976"}
+
+# ---------------------------------------------------------------------------
+# Top 10 cities by population (INSEE code → name)
+# ---------------------------------------------------------------------------
+TOP_10_CITIES: dict[str, str] = {
+    "75056": "Paris",
+    "13055": "Marseille",
+    "69123": "Lyon",
+    "31555": "Toulouse",
+    "06088": "Nice",
+    "44109": "Nantes",
+    "34172": "Montpellier",
+    "67482": "Strasbourg",
+    "33063": "Bordeaux",
+    "59350": "Lille",
+}
+
+# Paris, Lyon, Marseille have arrondissements — we need to map them back
+ARRONDISSEMENT_MAPPING: dict[str, str] = {}
+# Paris: 75101-75120 → 75056
+for i in range(1, 21):
+    ARRONDISSEMENT_MAPPING[f"751{i:02d}"] = "75056"
+# Lyon: 69381-69389 → 69123
+for i in range(1, 10):
+    ARRONDISSEMENT_MAPPING[f"6938{i}"] = "69123"
+# Marseille: 13201-13216 → 13055
+for i in range(1, 17):
+    ARRONDISSEMENT_MAPPING[f"132{i:02d}"] = "13055"
+
+# ---------------------------------------------------------------------------
+# Aggregation levels
+# ---------------------------------------------------------------------------
+AGGREGATION_LEVELS = [
+    "country",
+    "region",
+    "department",
+    "commune",
+    "postcode",
+    "section",
+]

src/downloader.py

@@ -0,0 +1,125 @@
+"""
+Download DVF geolocalized data from data.gouv.fr.
+
+Responsibility: Fetch raw CSV files and store them locally.
+Handles caching — won't re-download files that already exist.
+"""
+
+import gzip
+import logging
+from pathlib import Path
+
+import requests
+from tqdm import tqdm
+
+from src.config import DVF_YEARS, RAW_DIR, dvf_url
+
+logger = logging.getLogger(__name__)
+
+CHUNK_SIZE = 8192  # 8 KB chunks for streaming download
+
+
+def download_file(url: str, dest: Path, *, force: bool = False) -> Path:
+    """
+    Download a single file with progress bar and caching.
+
+    Args:
+        url:   Remote URL to download.
+        dest:  Local path to save the file.
+        force: If True, re-download even if file exists.
+
+    Returns:
+        Path to the downloaded file.
+
+    Raises:
+        requests.HTTPError: If the server returns a non-2xx status.
+    """
+    if dest.exists() and not force:
+        logger.info("Cached: %s", dest.name)
+        return dest
+
+    dest.parent.mkdir(parents=True, exist_ok=True)
+    logger.info("Downloading: %s", url)
+
+    response = requests.get(url, stream=True, timeout=300)
+    response.raise_for_status()
+
+    total = int(response.headers.get("content-length", 0))
+    with (
+        open(dest, "wb") as f,
+        tqdm(total=total, unit="B", unit_scale=True, desc=dest.name) as bar,
+    ):
+        for chunk in response.iter_content(chunk_size=CHUNK_SIZE):
+            f.write(chunk)
+            bar.update(len(chunk))
+
+    logger.info("Saved: %s (%.1f MB)", dest.name, dest.stat().st_size / 1e6)
+    return dest
+
+
+def decompress_gz(gz_path: Path) -> Path:
+    """
+    Decompress a .gz file to .csv in the same directory.
+
+    Args:
+        gz_path: Path to the .gz file.
+
+    Returns:
+        Path to the decompressed .csv file.
+    """
+    csv_path = gz_path.with_suffix("")  # removes .gz
+    if csv_path.exists():
+        logger.info("Already decompressed: %s", csv_path.name)
+        return csv_path
+
+    logger.info("Decompressing: %s", gz_path.name)
+    with gzip.open(gz_path, "rb") as f_in, open(csv_path, "wb") as f_out:
+        while chunk := f_in.read(CHUNK_SIZE * 128):
+            f_out.write(chunk)
+
+    logger.info("Decompressed: %s (%.1f MB)", csv_path.name, csv_path.stat().st_size / 1e6)
+    return csv_path
+
+
+def download_dvf_year(year: int, *, force: bool = False) -> Path:
+    """
+    Download and decompress DVF data for a single year.
+
+    Args:
+        year:  Year to download (e.g. 2024).
+        force: Re-download even if cached.
+
+    Returns:
+        Path to the decompressed CSV file.
+    """
+    url = dvf_url(year)
+    gz_path = RAW_DIR / f"dvf_{year}.csv.gz"
+    download_file(url, gz_path, force=force)
+    return decompress_gz(gz_path)
+
+
+def download_all(years: list[int] | None = None, *, force: bool = False) -> list[Path]:
+    """
+    Download DVF data for all configured years.
+
+    Args:
+        years: List of years to download. Defaults to config.DVF_YEARS.
+        force: Re-download even if cached.
+
+    Returns:
+        List of paths to decompressed CSV files.
+    """
+    years = years or DVF_YEARS
+    paths = []
+    for year in years:
+        try:
+            path = download_dvf_year(year, force=force)
+            paths.append(path)
+        except requests.HTTPError as e:
+            logger.error("Failed to download year %d: %s", year, e)
+    return paths
+
+
+if __name__ == "__main__":
+    logging.basicConfig(level=logging.INFO)
+    download_all()

src/pipeline.py

@@ -0,0 +1,110 @@
+"""
+Orchestrate the full DVF data pipeline.
+
+Responsibility: Wire together download → clean → aggregate → export.
+This is the main entry point for running the complete pipeline.
+"""
+
+import logging
+import time
+
+from src.config import AGGREGATED_DIR, PROCESSED_DIR, RAW_DIR
+from src.downloader import download_all
+from src.cleaner import clean
+from src.aggregator import aggregate_all_levels, export_json
+from src.top_cities import compute_top_cities, export_top_cities
+
+logger = logging.getLogger(__name__)
+
+
+def run(
+    *,
+    years: list[int] | None = None,
+    skip_download: bool = False,
+    skip_section: bool = False,
+) -> None:
+    """
+    Run the complete pipeline: download → clean → aggregate → export.
+
+    Args:
+        years:          Years to process. None = all configured years.
+        skip_download:  Skip download step (use existing raw files).
+        skip_section:   Skip section-level aggregation (slow, large output).
+    """
+    t0 = time.time()
+
+    # Step 1: Download
+    if not skip_download:
+        logger.info("=" * 60)
+        logger.info("STEP 1: Downloading DVF data")
+        logger.info("=" * 60)
+        csv_paths = download_all(years)
+    else:
+        logger.info("Skipping download, using existing files...")
+        csv_paths = sorted(RAW_DIR.glob("dvf_*.csv"))
+        csv_paths = [p for p in csv_paths if not p.name.endswith(".gz")]
+        logger.info("Found %d raw CSV files", len(csv_paths))
+
+    if not csv_paths:
+        logger.error("No CSV files found. Run without --skip-download first.")
+        return
+
+    # Step 2: Clean
+    logger.info("=" * 60)
+    logger.info("STEP 2: Cleaning data")
+    logger.info("=" * 60)
+    lf = clean(csv_paths)
+
+    # Materialize once and save as parquet for reuse
+    logger.info("Materializing cleaned data...")
+    PROCESSED_DIR.mkdir(parents=True, exist_ok=True)
+    df_clean = lf.collect()
+    parquet_path = PROCESSED_DIR / "dvf_clean.parquet"
+    df_clean.write_parquet(parquet_path)
+    logger.info(
+        "Saved: %s (%d rows, %.1f MB)",
+        parquet_path.name,
+        len(df_clean),
+        parquet_path.stat().st_size / 1e6,
+    )
+
+    # Step 3: Aggregate
+    logger.info("=" * 60)
+    logger.info("STEP 3: Aggregating prices")
+    logger.info("=" * 60)
+    lf_clean = df_clean.lazy()
+
+    if skip_section:
+        from src.config import AGGREGATION_LEVELS
+        levels_to_run = [l for l in AGGREGATION_LEVELS if l != "section"]
+        from src.aggregator import LEVEL_TO_COLUMN, aggregate_all_types
+        aggregated = {}
+        for level in levels_to_run:
+            col = LEVEL_TO_COLUMN[level]
+            logger.info("Aggregating: %s", level)
+            aggregated[level] = aggregate_all_types(lf_clean, col)
+    else:
+        aggregated = aggregate_all_levels(lf_clean)
+
+    export_json(aggregated)
+
+    # Step 4: Top cities
+    logger.info("=" * 60)
+    logger.info("STEP 4: Top 10 cities")
+    logger.info("=" * 60)
+    df_cities = compute_top_cities(lf_clean)
+    export_top_cities(df_cities)
+
+    elapsed = time.time() - t0
+    logger.info("=" * 60)
+    logger.info("Pipeline complete in %.1f seconds", elapsed)
+    logger.info("Output: %s", AGGREGATED_DIR)
+    logger.info("=" * 60)
+
+
+if __name__ == "__main__":
+    logging.basicConfig(
+        level=logging.INFO,
+        format="%(asctime)s %(levelname)s %(name)s: %(message)s",
+    )
+    run()

src/top_cities.py

@@ -0,0 +1,121 @@
+"""
+Compute price/m² breakdown for the top 10 French cities.
+
+Responsibility: Produce a clean table of median price per m²
+by property type for the largest cities.
+"""
+
+import json
+import logging
+from pathlib import Path
+
+import polars as pl
+
+from src.config import AGGREGATED_DIR, TOP_10_CITIES, TYPE_LOCAL_SHORT
+
+logger = logging.getLogger(__name__)
+
+
+def compute_top_cities(lf: pl.LazyFrame) -> pl.DataFrame:
+    """
+    Compute price statistics for top 10 cities, broken down by property type.
+
+    Uses code_commune_city (with arrondissements mapped to parent city)
+    to correctly aggregate Paris, Lyon, and Marseille.
+
+    Args:
+        lf: Cleaned LazyFrame with code_commune_city column.
+
+    Returns:
+        DataFrame with columns: city_code, city_name, type, median_price_m2,
+        mean_price_m2, volume, q1, q3.
+    """
+    city_codes = list(TOP_10_CITIES.keys())
+
+    # Filter to top 10 cities only
+    city_data = lf.filter(
+        pl.col("code_commune_city").is_in(city_codes)
+    )
+
+    results = []
+
+    # All types combined
+    df_all = (
+        city_data.group_by("code_commune_city")
+        .agg(
+            pl.col("prix_m2").median().alias("median_price_m2"),
+            pl.col("prix_m2").mean().alias("mean_price_m2"),
+            pl.col("prix_m2").quantile(0.25).alias("q1"),
+            pl.col("prix_m2").quantile(0.75).alias("q3"),
+            pl.col("prix_m2").count().alias("volume"),
+        )
+        .with_columns(pl.lit("tous").alias("type"))
+        .collect()
+    )
+    results.append(df_all)
+
+    # Per property type
+    for full_name, short_name in TYPE_LOCAL_SHORT.items():
+        df_type = (
+            city_data.filter(pl.col("type_local") == full_name)
+            .group_by("code_commune_city")
+            .agg(
+                pl.col("prix_m2").median().alias("median_price_m2"),
+                pl.col("prix_m2").mean().alias("mean_price_m2"),
+                pl.col("prix_m2").quantile(0.25).alias("q1"),
+                pl.col("prix_m2").quantile(0.75).alias("q3"),
+                pl.col("prix_m2").count().alias("volume"),
+            )
+            .with_columns(pl.lit(short_name).alias("type"))
+            .collect()
+        )
+        results.append(df_type)
+
+    combined = pl.concat(results, how="vertical_relaxed")
+
+    # Add city names
+    city_name_map = {code: name for code, name in TOP_10_CITIES.items()}
+    combined = combined.with_columns(
+        pl.col("code_commune_city")
+        .replace_strict(city_name_map, default="Unknown")
+        .alias("city_name")
+    ).rename({"code_commune_city": "city_code"})
+
+    return combined.sort(["city_name", "type"])
+
+
+def export_top_cities(df: pl.DataFrame, output_dir: Path | None = None) -> None:
+    """
+    Export top cities data to JSON.
+
+    Output format:
+    {
+        "Paris": {
+            "code": "75056",
+            "tous": {"median": 10500, "volume": 45000, ...},
+            "appartement": {...},
+            ...
+        },
+        ...
+    }
+    """
+    output_dir = output_dir or AGGREGATED_DIR
+    output_dir.mkdir(parents=True, exist_ok=True)
+
+    data: dict = {}
+    for row in df.iter_rows(named=True):
+        name = row["city_name"]
+        if name not in data:
+            data[name] = {"code": row["city_code"]}
+        data[name][row["type"]] = {
+            "median": round(row["median_price_m2"] or 0, 1),
+            "mean": round(row["mean_price_m2"] or 0, 1),
+            "q1": round(row["q1"] or 0, 1),
+            "q3": round(row["q3"] or 0, 1),
+            "volume": row["volume"],
+        }
+
+    path = output_dir / "top_cities.json"
+    with open(path, "w") as f:
+        json.dump(data, f, ensure_ascii=False, indent=2)
+    logger.info("Exported: %s (%d cities)", path.name, len(data))