Model Training and Evaluation (Myanmar)

import os
import sys
sys.path.append("../../")

import uuid
import numpy as np
import geopandas as gpd
import pandas as pd
from shapely import wkt

from geowrangler import dhs
from povertymapping import settings, osm, ookla, nightlights
from povertymapping.dhs import generate_dhs_cluster_level_data
from povertymapping.osm import OsmDataManager
from povertymapping.ookla import OoklaDataManager
import getpass

import pickle


from sklearn.model_selection import train_test_split, KFold, RepeatedKFold
from sklearn.model_selection import GroupKFold, cross_val_predict, cross_val_score
from sklearn.metrics import r2_score
import matplotlib.pyplot as plt
import seaborn as sns

import shap

/home/jc_tm/project_repos/unicef-ai4d-poverty-mapping/env/lib/python3.9/site-packages/geopandas/_compat.py:111: UserWarning: The Shapely GEOS version (3.10.3-CAPI-1.16.1) is incompatible with the GEOS version PyGEOS was compiled with (3.10.1-CAPI-1.16.0). Conversions between both will be slow.
  warnings.warn(
/home/jc_tm/project_repos/unicef-ai4d-poverty-mapping/env/lib/python3.9/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html
  from .autonotebook import tqdm as notebook_tqdm

%reload_ext autoreload
%autoreload 2

Load Target Country From DHS data

# Set country-specific variables
country_osm = "myanmar"
ookla_year = 2019
nightlights_year = 2015

dhs_household_dta_path = settings.DATA_DIR/"dhs/mm/MMHR71DT/MMHR71FL.DTA"
dhs_geographic_shp_path = settings.DATA_DIR/"dhs/mm/MMGE71FL/MMGE71FL.shp"
dhs_gdf = generate_dhs_cluster_level_data(
    dhs_household_dta_path, 
    dhs_geographic_shp_path, 
    col_rename_config="mm",
    convert_geoms_to_bbox=True,
    bbox_size_km=2.4
).reset_index(drop=True)
dhs_gdf.explore()

Make this Notebook Trusted to load map: File -> Trust Notebook

dhs_gdf.head()

	DHSCLUST	Wealth Index	DHSID	DHSCC	DHSYEAR	CCFIPS	ADM1FIPS	ADM1FIPSNA	ADM1SALBNA	ADM1SALBCO	...	DHSREGCO	DHSREGNA	SOURCE	URBAN_RURA	LATNUM	LONGNUM	ALT_GPS	ALT_DEM	DATUM	geometry
0	1	-52232.000000	MM201500000001	MM	2015.0	BM	NULL	NULL	NULL	NULL	...	8.0	Magway	GPS	R	20.058637	95.360081	107.0	105.0	WGS84	POLYGON ((95.34859 20.06943, 95.37157 20.06943...
1	2	130773.724138	MM201500000002	MM	2015.0	BM	NULL	NULL	NULL	NULL	...	12.0	Yangon	GPS	U	17.112398	96.045616	31.0	14.0	WGS84	POLYGON ((96.03432 17.12319, 96.05691 17.12319...
2	3	-4955.000000	MM201500000003	MM	2015.0	BM	NULL	NULL	NULL	NULL	...	10.0	Mon	GPS	R	16.507664	97.364236	4.0	2.0	WGS84	POLYGON ((97.35298 16.51846, 97.37549 16.51846...
3	4	47824.103448	MM201500000004	MM	2015.0	BM	NULL	NULL	NULL	NULL	...	1.0	Kachin	GPS	U	26.684519	96.283879	193.0	210.0	WGS84	POLYGON ((96.27180 26.69531, 96.29596 26.69531...
4	5	9434.482759	MM201500000005	MM	2015.0	BM	NULL	NULL	NULL	NULL	...	12.0	Yangon	GPS	R	16.866059	96.053499	12.0	5.0	WGS84	POLYGON ((96.04222 16.87685, 96.06478 16.87685...

5 rows × 22 columns

Set up Data Access

# Instantiate data managers for Ookla and OSM
# This auto-caches requested data in RAM, so next fetches of the data are faster.
osm_data_manager = OsmDataManager(cache_dir=settings.ROOT_DIR/"data/data_cache")
ookla_data_manager = OoklaDataManager(cache_dir=settings.ROOT_DIR/"data/data_cache")

# Log-in using EOG credentials
username = os.environ.get('EOG_USER',None)
username = username if username is not None else input('Username?')
password = os.environ.get('EOG_PASSWORD',None)
password = password if password is not None else getpass.getpass('Password?') 


# set save_token to True so that access token gets stored in ~/.eog_creds/eog_access_token
access_token = nightlights.get_eog_access_token(username,password, save_token=True)

2023-01-31 13:58:23.564 | INFO     | povertymapping.nightlights:get_eog_access_token:48 - Saving access_token to ~/.eog_creds/eog_access_token
2023-01-31 13:58:23.570 | INFO     | povertymapping.nightlights:get_eog_access_token:56 - Adding access token to environmentt var EOG_ACCESS_TOKEN

Generate Base Features

If this is your first time running this notebook for this specific area, expect a long runtime for the following cell as it will download and cache the ff. datasets from the internet.

• OpenStreetMap Data from Geofabrik
• Ookla Internet Speed Data
• VIIRS nighttime lights data from NASA EOG

On subsequent runs, the runtime will be much faster as the data is already stored in your filesystem.

%%time
country_data = dhs_gdf.copy()

# Add in OSM features
country_data = osm.add_osm_poi_features(country_data, country_osm, osm_data_manager)
country_data = osm.add_osm_road_features(country_data, country_osm, osm_data_manager)

# Add in Ookla features
country_data = ookla.add_ookla_features(country_data, 'fixed', ookla_year, ookla_data_manager)
country_data = ookla.add_ookla_features(country_data, 'mobile', ookla_year, ookla_data_manager)

# Add in the nighttime lights features
country_data = nightlights.generate_nightlights_feature(country_data, str(nightlights_year)) 

2023-01-31 13:58:23.766 | INFO     | povertymapping.osm:download_osm_country_data:187 - OSM Data: Cached data available for myanmar at /home/jc_tm/project_repos/unicef-ai4d-poverty-mapping/notebooks/2023-01-17-initial-model-ph-mm-tl-kh/../../data/data_cache/osm/myanmar? True
2023-01-31 13:58:23.767 | DEBUG    | povertymapping.osm:load_pois:149 - OSM POIs for myanmar being loaded from /home/jc_tm/project_repos/unicef-ai4d-poverty-mapping/notebooks/2023-01-17-initial-model-ph-mm-tl-kh/../../data/data_cache/osm/myanmar/gis_osm_pois_free_1.shp
2023-01-31 13:58:33.290 | INFO     | povertymapping.osm:download_osm_country_data:187 - OSM Data: Cached data available for myanmar at /home/jc_tm/project_repos/unicef-ai4d-poverty-mapping/notebooks/2023-01-17-initial-model-ph-mm-tl-kh/../../data/data_cache/osm/myanmar? True
2023-01-31 13:58:33.297 | DEBUG    | povertymapping.osm:load_roads:168 - OSM Roads for myanmar being loaded from /home/jc_tm/project_repos/unicef-ai4d-poverty-mapping/notebooks/2023-01-17-initial-model-ph-mm-tl-kh/../../data/data_cache/osm/myanmar/gis_osm_roads_free_1.shp
2023-01-31 14:00:36.799 | DEBUG    | povertymapping.ookla:load_type_year_data:68 - Contents of data cache: []
2023-01-31 14:00:36.802 | INFO     | povertymapping.ookla:load_type_year_data:83 - Cached data available at /home/jc_tm/project_repos/unicef-ai4d-poverty-mapping/notebooks/2023-01-17-initial-model-ph-mm-tl-kh/../../data/data_cache/ookla/processed/d72ec7e4d144b750e1c0950ecad081e0.csv? True
2023-01-31 14:00:36.803 | DEBUG    | povertymapping.ookla:load_type_year_data:88 - Processed Ookla data for aoi, fixed 2019 (key: d72ec7e4d144b750e1c0950ecad081e0) found in filesystem. Loading in cache.
2023-01-31 14:00:37.049 | DEBUG    | povertymapping.ookla:load_type_year_data:68 - Contents of data cache: ['d72ec7e4d144b750e1c0950ecad081e0']
2023-01-31 14:00:37.050 | INFO     | povertymapping.ookla:load_type_year_data:83 - Cached data available at /home/jc_tm/project_repos/unicef-ai4d-poverty-mapping/notebooks/2023-01-17-initial-model-ph-mm-tl-kh/../../data/data_cache/ookla/processed/2aff65fdf8072457cba0d42873b7a9c2.csv? True
2023-01-31 14:00:37.051 | DEBUG    | povertymapping.ookla:load_type_year_data:88 - Processed Ookla data for aoi, mobile 2019 (key: 2aff65fdf8072457cba0d42873b7a9c2) found in filesystem. Loading in cache.
2023-01-31 14:00:37.265 | INFO     | povertymapping.nightlights:get_clipped_raster:414 - Retrieving clipped raster file /home/jc_tm/.geowrangler/nightlights/clip/7a58f067614b6685cd9bb62d4d15a249.tif

CPU times: user 1min 40s, sys: 19.6 s, total: 2min
Wall time: 2min 16s

Inspect the combined target country data

country_data.info()

<class 'geopandas.geodataframe.GeoDataFrame'>
Int64Index: 441 entries, 0 to 440
Data columns (total 83 columns):
 #   Column                             Non-Null Count  Dtype   
---  ------                             --------------  -----   
 0   DHSCLUST                           441 non-null    int64   
 1   Wealth Index                       441 non-null    float64 
 2   DHSID                              441 non-null    object  
 3   DHSCC                              441 non-null    object  
 4   DHSYEAR                            441 non-null    float64 
 5   CCFIPS                             441 non-null    object  
 6   ADM1FIPS                           441 non-null    object  
 7   ADM1FIPSNA                         441 non-null    object  
 8   ADM1SALBNA                         441 non-null    object  
 9   ADM1SALBCO                         441 non-null    object  
 10  ADM1DHS                            441 non-null    float64 
 11  ADM1NAME                           441 non-null    object  
 12  DHSREGCO                           441 non-null    float64 
 13  DHSREGNA                           441 non-null    object  
 14  SOURCE                             441 non-null    object  
 15  URBAN_RURA                         441 non-null    object  
 16  LATNUM                             441 non-null    float64 
 17  LONGNUM                            441 non-null    float64 
 18  ALT_GPS                            441 non-null    float64 
 19  ALT_DEM                            441 non-null    float64 
 20  DATUM                              441 non-null    object  
 21  geometry                           441 non-null    geometry
 22  poi_count                          441 non-null    float64 
 23  atm_count                          441 non-null    float64 
 24  atm_nearest                        441 non-null    float64 
 25  bank_count                         441 non-null    float64 
 26  bank_nearest                       441 non-null    float64 
 27  bus_station_count                  441 non-null    float64 
 28  bus_station_nearest                441 non-null    float64 
 29  cafe_count                         441 non-null    float64 
 30  cafe_nearest                       441 non-null    float64 
 31  charging_station_count             441 non-null    float64 
 32  charging_station_nearest           441 non-null    float64 
 33  courthouse_count                   441 non-null    float64 
 34  courthouse_nearest                 441 non-null    float64 
 35  dentist_count                      441 non-null    float64 
 36  dentist_nearest                    441 non-null    float64 
 37  fast_food_count                    441 non-null    float64 
 38  fast_food_nearest                  441 non-null    float64 
 39  fire_station_count                 441 non-null    float64 
 40  fire_station_nearest               441 non-null    float64 
 41  food_court_count                   441 non-null    float64 
 42  food_court_nearest                 441 non-null    float64 
 43  fuel_count                         441 non-null    float64 
 44  fuel_nearest                       441 non-null    float64 
 45  hospital_count                     441 non-null    float64 
 46  hospital_nearest                   441 non-null    float64 
 47  library_count                      441 non-null    float64 
 48  library_nearest                    441 non-null    float64 
 49  marketplace_count                  441 non-null    float64 
 50  marketplace_nearest                441 non-null    float64 
 51  pharmacy_count                     441 non-null    float64 
 52  pharmacy_nearest                   441 non-null    float64 
 53  police_count                       441 non-null    float64 
 54  police_nearest                     441 non-null    float64 
 55  post_box_count                     441 non-null    float64 
 56  post_box_nearest                   441 non-null    float64 
 57  post_office_count                  441 non-null    float64 
 58  post_office_nearest                441 non-null    float64 
 59  restaurant_count                   441 non-null    float64 
 60  restaurant_nearest                 441 non-null    float64 
 61  social_facility_count              441 non-null    float64 
 62  social_facility_nearest            441 non-null    float64 
 63  supermarket_count                  441 non-null    float64 
 64  supermarket_nearest                441 non-null    float64 
 65  townhall_count                     441 non-null    float64 
 66  townhall_nearest                   441 non-null    float64 
 67  road_count                         441 non-null    float64 
 68  fixed_2019_mean_avg_d_kbps_mean    149 non-null    float64 
 69  fixed_2019_mean_avg_u_kbps_mean    149 non-null    float64 
 70  fixed_2019_mean_avg_lat_ms_mean    149 non-null    float64 
 71  fixed_2019_mean_num_tests_mean     149 non-null    float64 
 72  fixed_2019_mean_num_devices_mean   149 non-null    float64 
 73  mobile_2019_mean_avg_d_kbps_mean   247 non-null    float64 
 74  mobile_2019_mean_avg_u_kbps_mean   247 non-null    float64 
 75  mobile_2019_mean_avg_lat_ms_mean   247 non-null    float64 
 76  mobile_2019_mean_num_tests_mean    247 non-null    float64 
 77  mobile_2019_mean_num_devices_mean  247 non-null    float64 
 78  avg_rad_min                        441 non-null    float64 
 79  avg_rad_max                        441 non-null    float64 
 80  avg_rad_mean                       441 non-null    float64 
 81  avg_rad_std                        441 non-null    float64 
 82  avg_rad_median                     441 non-null    float64 
dtypes: float64(69), geometry(1), int64(1), object(12)
memory usage: 305.6+ KB

country_data.head()

	DHSCLUST	Wealth Index	DHSID	DHSCC	DHSYEAR	CCFIPS	ADM1FIPS	ADM1FIPSNA	ADM1SALBNA	ADM1SALBCO	...	mobile_2019_mean_avg_d_kbps_mean	mobile_2019_mean_avg_u_kbps_mean	mobile_2019_mean_avg_lat_ms_mean	mobile_2019_mean_num_tests_mean	mobile_2019_mean_num_devices_mean	avg_rad_min	avg_rad_max	avg_rad_mean	avg_rad_std	avg_rad_median
0	1	-52232.000000	MM201500000001	MM	2015.0	BM	NULL	NULL	NULL	NULL	...	NaN	NaN	NaN	NaN	NaN	0.084249	0.116014	0.100290	0.008462	0.100869
1	2	130773.724138	MM201500000002	MM	2015.0	BM	NULL	NULL	NULL	NULL	...	826.298933	505.281935	1.226730	0.166726	0.077713	0.413245	3.301661	1.938653	0.959708	1.911360
2	3	-4955.000000	MM201500000003	MM	2015.0	BM	NULL	NULL	NULL	NULL	...	843.116769	375.838860	1.581117	0.281163	0.082694	0.075951	1.287470	0.431493	0.367520	0.277878
3	4	47824.103448	MM201500000004	MM	2015.0	BM	NULL	NULL	NULL	NULL	...	NaN	NaN	NaN	NaN	NaN	0.087990	0.145543	0.117616	0.015429	0.115754
4	5	9434.482759	MM201500000005	MM	2015.0	BM	NULL	NULL	NULL	NULL	...	745.823942	688.509072	0.889722	0.244145	0.141459	1.886424	13.847748	9.098659	2.738587	9.385050

5 rows × 83 columns

Data Preparation

Split into labels and features

# Set parameters
label_col = 'Wealth Index'

# Split train/test data into features and labels

# For labels, we just select the target label column
labels = country_data[[label_col]]

# For features, drop all columns from the input country geometries
# If you need the cluster data, refer to country_data / country_test
input_dhs_cols = dhs_gdf.columns
features = country_data.drop(input_dhs_cols, axis=1)

features.shape, labels.shape

((441, 61), (441, 1))

# Clean features
# For now, just impute nans with 0
# TODO: Implement other cleaning steps
features = features.fillna(0)

Base Features List

The features can be subdivided by the source dataset

OSM

• <poi type>_count: number of points of interest (POI) of a specified type in that area
  • ex. atm_count: number of atms in cluster
  • poi_count: number of all POIs of all types in cluster
• <poi_type>_nearest: distance of nearest POI of the specified type
  • ex. atm_nearest: distance of nearest ATM from that cluster
• OSM POI types included: atm, bank, bus_stations, cafe, charging_station, courthouse, dentist (clinic), fast_food, fire_station, food_court, fuel (gas station), hospital, library, marketplace, pharmacy, police, post_box, post_office, restaurant, social_facility, supermarket, townhall, road

Ookla

The network metrics features follow the following name convention:

<type>_<year>_<yearly aggregate>_<network variable>_<cluster aggregate>

• type: kind of network connection measured
  • fixed: connection from fixed sources (landline, fiber, etc.)
  • mobile: connection from mobile devices
• year: Year of source data
• yearly aggregate: How data was aggregated into yearly data
  • Note: Ookla provides data per quarter, so a yearly mean takes the average across 4 quarters
  • For this model, we only aggregate by yearly mean
• network variable: network characteristic described
  • avg_d_kbps: average download speed in kbps
  • avg_u_kbps: average upload speed in kbps
  • avg_lat_ms: average latency in ms
  • num_devices: number of devices measured
• cluster aggregate: how the data was aggregated per cluster aggregate
  • Types: min, mean, max, median, std.
    • For this model: only mean is used
  • This is calculated using area zonal stats, which weighs the average by the intersection of the Ookla tile with the cluster geometry.

Ex. fixed_2019_mean_avg_d_kbps_median takes the cluster median of 2019 yearly average download speed.

Nightlights (VIIRS)

All nightlights features are taken as the zonal aggregate of the raster data per cluster

• ex. avg_rad_mean: cluster mean of the average radiance
• aggregations used: min, mean, max, median

features.info()

<class 'geopandas.geodataframe.GeoDataFrame'>
Int64Index: 441 entries, 0 to 440
Data columns (total 61 columns):
 #   Column                             Non-Null Count  Dtype  
---  ------                             --------------  -----  
 0   poi_count                          441 non-null    float64
 1   atm_count                          441 non-null    float64
 2   atm_nearest                        441 non-null    float64
 3   bank_count                         441 non-null    float64
 4   bank_nearest                       441 non-null    float64
 5   bus_station_count                  441 non-null    float64
 6   bus_station_nearest                441 non-null    float64
 7   cafe_count                         441 non-null    float64
 8   cafe_nearest                       441 non-null    float64
 9   charging_station_count             441 non-null    float64
 10  charging_station_nearest           441 non-null    float64
 11  courthouse_count                   441 non-null    float64
 12  courthouse_nearest                 441 non-null    float64
 13  dentist_count                      441 non-null    float64
 14  dentist_nearest                    441 non-null    float64
 15  fast_food_count                    441 non-null    float64
 16  fast_food_nearest                  441 non-null    float64
 17  fire_station_count                 441 non-null    float64
 18  fire_station_nearest               441 non-null    float64
 19  food_court_count                   441 non-null    float64
 20  food_court_nearest                 441 non-null    float64
 21  fuel_count                         441 non-null    float64
 22  fuel_nearest                       441 non-null    float64
 23  hospital_count                     441 non-null    float64
 24  hospital_nearest                   441 non-null    float64
 25  library_count                      441 non-null    float64
 26  library_nearest                    441 non-null    float64
 27  marketplace_count                  441 non-null    float64
 28  marketplace_nearest                441 non-null    float64
 29  pharmacy_count                     441 non-null    float64
 30  pharmacy_nearest                   441 non-null    float64
 31  police_count                       441 non-null    float64
 32  police_nearest                     441 non-null    float64
 33  post_box_count                     441 non-null    float64
 34  post_box_nearest                   441 non-null    float64
 35  post_office_count                  441 non-null    float64
 36  post_office_nearest                441 non-null    float64
 37  restaurant_count                   441 non-null    float64
 38  restaurant_nearest                 441 non-null    float64
 39  social_facility_count              441 non-null    float64
 40  social_facility_nearest            441 non-null    float64
 41  supermarket_count                  441 non-null    float64
 42  supermarket_nearest                441 non-null    float64
 43  townhall_count                     441 non-null    float64
 44  townhall_nearest                   441 non-null    float64
 45  road_count                         441 non-null    float64
 46  fixed_2019_mean_avg_d_kbps_mean    441 non-null    float64
 47  fixed_2019_mean_avg_u_kbps_mean    441 non-null    float64
 48  fixed_2019_mean_avg_lat_ms_mean    441 non-null    float64
 49  fixed_2019_mean_num_tests_mean     441 non-null    float64
 50  fixed_2019_mean_num_devices_mean   441 non-null    float64
 51  mobile_2019_mean_avg_d_kbps_mean   441 non-null    float64
 52  mobile_2019_mean_avg_u_kbps_mean   441 non-null    float64
 53  mobile_2019_mean_avg_lat_ms_mean   441 non-null    float64
 54  mobile_2019_mean_num_tests_mean    441 non-null    float64
 55  mobile_2019_mean_num_devices_mean  441 non-null    float64
 56  avg_rad_min                        441 non-null    float64
 57  avg_rad_max                        441 non-null    float64
 58  avg_rad_mean                       441 non-null    float64
 59  avg_rad_std                        441 non-null    float64
 60  avg_rad_median                     441 non-null    float64
dtypes: float64(61)
memory usage: 229.8 KB

Model Training

# Set parameters
cv_col = 'ADM1NAME'
cv_num_splits = 5
cv_num_repeats = 5
train_test_seed = 42
test_size = 0.2

Create train/test cross-validation indices

# train_features, test_features, train_labels, test_labels = train_test_split(
#     features, labels, test_size=test_size, random_state=train_test_seed
# )

# Cross validation
print(f"Performing {cv_num_splits}-fold CV...")
cv = RepeatedKFold(n_splits=cv_num_splits, n_repeats=cv_num_repeats, random_state=train_test_seed)

print(cv.split(features))

Performing 5-fold CV...
<generator object _RepeatedSplits.split at 0x7eff0a1c7040>

Instantiate model

For now, we will train a simple random forest model

from sklearn.ensemble import RandomForestRegressor
model = RandomForestRegressor(n_estimators=100, random_state=train_test_seed, verbose=0)
model

RandomForestRegressor(random_state=42)

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

Evaluate model training using cross-validation

We evalute the model’s generalizability when training over different train/test splits

Ideally for R^2 - We want a high mean: This means that we achieve a high model performance over the different train/test splits - We want a low standard deviation (std): This means that the model performance is stable over multiple training repetitions

R_cv = cross_val_score(model, features.values, labels.values.ravel(), cv=cv)
print("Cross validation scores are: ", R_cv)
cv_mean = round(np.array(R_cv).mean(), 2)
cv_std = round(np.array(R_cv).std(), 2)
print(f"Cross validation R^2 mean: {cv_mean}")
print(f"Cross validation R^2 std: {cv_std}")

Cross validation scores are:  [0.49816309 0.52690891 0.36088424 0.5387295  0.55555747 0.45891165
 0.51177753 0.50802007 0.61784092 0.24206319 0.48448388 0.65377474
 0.44347434 0.46700331 0.52625576 0.50967757 0.59711872 0.34610694
 0.47138862 0.60987113 0.52611774 0.43326341 0.38955954 0.53080727
 0.54003998]
Cross validation R^2 mean: 0.49
Cross validation R^2 std: 0.09

Train the final model

For training the final model, we train on all the available data.

model.fit(features.values, labels.values.ravel())

RandomForestRegressor(random_state=42)

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

Model Evaluation

SHAP Feature Importance

explainer = shap.TreeExplainer(model)
shap_values = explainer.shap_values(features)

shap_values

array([[-1208.78718752,   -89.14060743,  -933.90599278, ...,
        -2406.63320409, -5513.91433576, -6400.69078997],
       [ 3906.12555776,  -259.13018384,  -686.35617074, ...,
         1394.55206468,  2252.46954147,  3214.22551448],
       [ 2542.24891001,   -81.73893167, -1169.67173119, ...,
         1702.14606374,   891.70030008,  2319.03870166],
       ...,
       [-3738.33959336,  -273.39904489, -2169.94002487, ...,
          570.19299015,  1908.05455596,   399.26280909],
       [ -991.90362228,   -86.45828226, -1092.65620164, ...,
         4285.58797922,    40.72512533,  9393.94536717],
       [ 6149.21300248,  2420.50197243,  3649.90180698, ...,
         3886.7001019 ,   -51.63548071,  4208.20211146]])

shap.summary_plot(shap_values, features, feature_names=features.columns, plot_type="bar")

shap.summary_plot(shap_values, features.values, feature_names=features.columns)

No data for colormapping provided via 'c'. Parameters 'vmin', 'vmax' will be ignored

Save Model

model_save_path = "./model_mm.pkl"
with open(model_save_path, "wb") as file:
    pickle.dump(model, file)