geomapviz 0.4

Geographical visualization

buy me caffeine

Geographical Visualization

Geomapviz is a package to visualize data on maps at different levels of granularity, aggregating at the specified geographical level, dissolving the polygons if needed. The maps can be static (using matplotlib) or interactive (using geoviews and holoviews).

It returns a single map or a panel of maps, useful to compare how different models are capturing the geographical patterns.

The average values are either returned raw or (auto) binned. You can change the face-colour (background) if you prefer light over a black background (I don't know if anyone would).

Installation

$ pip install geomapviz -U

Applications

Generate dummy data

Introducing correlation to the geo identifier of regions

import geomapviz as gm
import cmasher as cmr

# the greatest country in the world,
# first military and economic power in the Universe
shp_file = gm.load_be_shp()
geom_df = shp_file.copy()

# create correlation with the geo entities
feat_1 = np.repeat(np.log10(geom_df.INS.astype(int).values), 10)
feat_1 = (feat_1 - feat_1.min()) / (feat_1.max() - feat_1.min())
# dummy data
X = (np.repeat(geom_df.long.values, 10) - (geom_df.long.mean())) / geom_df.long.std()
Y = (np.repeat(geom_df.lat.values, 10) - (geom_df.lat.mean())) / geom_df.lat.std()

# dummy data
bel_df = pd.DataFrame({
    'geoid': np.repeat(geom_df.INS.values, 10),
    'truth': (1 - Y + X + Y**3) * np.exp(-(X**2 + Y**2)),
    'feat_2':  (1 - Y**3 + X**3 + Y**5) * np.exp(-(X**2 + Y**2)) + np.random.beta(.5, .5, size=len(feat_1)),
    'feat_3': (1 + Y*X+ Y**3) * np.exp(-(X**2 + Y**2)) + np.random.beta(.5, .5, size=len(feat_1)),
    'feat_4': feat_1 + np.random.beta(5, 2, size=len(feat_1))
}
)

bel_df = bel_df.merge(geom_df[['INS', 'borough', 'district']], left_on='geoid', right_on='INS')

Simple choropleth

f = gm.plot_on_map(df=bel_df, target='truth', geoid='INS', shp_file=shp_file,
                   figsize=(20, 6), cmap=cmr.iceburn, normalize=True, facecolor="black")

Simple choropleth and incertitude on the average

f = gm.plot_on_map(df=bel_df, target='truth', dissolve_on=None, distrib='gaussian',
                   plot_uncertainty=True, plot_weight=False,
                   autobin=False, n_bins=7, geoid='INS', weight=None, shp_file=shp_file,
                   figsize=(20, 6), cmap=cmr.iceburn, normalize=True, facecolor="black")

Simple choropleth with auto-binning

You can also choose the number of decimal in the legend.

f = gm.plot_on_map(df=bel_df, target='truth', plot_uncertainty=True,
                   autobin=True, n_bins=7, geoid='INS', weight=None, shp_file=shp_file,
                   figsize=(20, 6), cmap=cmr.iceburn, normalize=True, facecolor="black", nbr_of_dec=3)

Panel of choropleths with auto-binning and dissolve polygons

cols_pred = ['feat_2', 'feat_3', 'feat_4']
f = gm.facet_map(df=bel_df, target='truth', cols_to_plot=cols_pred, dissolve_on='borough',
                      autobin=True, n_bins=5, geoid='INS', shp_file=shp_file
                      figsize=(12, 12), ncols=2, normalize=False)

Interactive map

Using geoviews behind the scene, geomapviz generates an interactive map. You can have non-binned or auto-binned average values, etc.

f = gm.facet_map_interactive(df=bel_df, target='truth', cols_to_plot=None, predicted=None, dissolve_on=None,
                             autobin=True, n_bins=7, geoid='INS', weight=None, shp_file=shp_file, alpha=.8,
                             figsize=(400, 400), ncols=2, cmap=cmr.iceburn, normalize=False) #, tiles_src='Wikipedia')
f

Interactive panel

By passing column names , you will have an interactive panel chart with linked maps. You can change the source of the tiles if you want to, using the tiles_src argument.

cols_pred = ['feat_2', 'feat_3', 'feat_4']
f = gm.facet_map_interactive(df=bel_df, target='truth', cols_to_plot=cols_pred, predicted=None, dissolve_on=None,
                             autobin=True, n_bins=7, geoid='INS', weight=None, shp_file=shp_file,
                             figsize=(400, 400), ncols=2, cmap=None, normalize=False) #, tiles_src='Wikipedia')
f

Changes

0.4

• Make Belgian shp available using load_be_shp
• More decimal
• User defined alpha for the interactive maps

0.3

• Bound functions to the upper level

0.2

• First version

0.1

• First version