def merge_map_data(path, featuredf, store=False):
"""
For each set of features, take results of graph reduction,
assign cluster numbers, and compute all other data for insertion
into final geojson file.
Args:
path : path to directory containing graph cut list files, string
featuredf : scaled features, pandas dataframe
Returns:
Data for insertion into geojson, pandas dataframe
"""
# get filenames
files = os.listdir(path)
files = [f[2:-4] for f in files if f[:2] == 'CL']
# null map
files.remove('xx')
# only allow 3 or less features
mapnos = [f for f in files if len(f) <= 6]
fnums = [mapno2list(f) for f in mapnos]
# column names
fnames = map(lambda x: [FDICT[n] for n in x], fnums)
# fixed number of clusters
nclustersmax = 28
# make null map
cnum = cut2cluster('xx', nclustersmax, allowed_nodes=featuredf.index)
# retain only mutual nodes
nodelist = set(featuredf.index).intersection(set(cnum.index))
featuredf = featuredf.ix[nodelist]
cnum = cnum.ix[nodelist]
nclusters = len(cnum.unique())
# compute data
# similarity colors
rgmatrix = rg_colormatrix(most_similar(featuredf, cnum))
# feature bar graph data
fbars = feature_bars(featuredf[FDICT.values()], cnum)
# shape file polygons
fn = 'data/uscensus/tl_2010_06075_tabblock10/tl_2010_06075_tabblock10.dbf'
mergedf = merge_shapefiles(featuredf[['lat', 'lon']], fn)
polys = make_shapefiles(featuredf[['lat', 'lon']], mergedf.polys, cnum)
# compile into single dataframe
alldf = pd.DataFrame({'cnum': cnum.unique(),
'polygon': polys})
alldf['rgmatrix'] = map(lambda x: list(rgmatrix.ix[x]), cnum.unique())
alldf['mapno'] = ''
alldf['fbars'] = map(list, fbars.round(2).values)
# store results
if store:
alldf.to_csv('results/geojson.csv')
# make all other maps
for i, f in enumerate(mapnos):
cnum = cut2cluster(f, nclustersmax, allowed_nodes=featuredf.index)
rgmatrix = rg_colormatrix(most_similar(featuredf, cnum))
fbars = feature_bars(featuredf[fnames[i]], cnum)
polys = make_shapefiles(featuredf[['lat', 'lon']],
mergedf.polys, cnum)
onedf = pd.DataFrame({'cnum': cnum.unique(),
'polygon': polys})
onedf['rgmatrix'] = map(lambda x: list(rgmatrix.ix[x]), cnum.unique())
onedf['mapno'] = f
onedf['fbars'] = map(list, fbars.round(2).values)
# append results after each map
if store:
with open('results/geojson.csv', 'a') as storefile:
onedf.to_csv(storefile, header=False)
alldf = pd.concat((alldf, onedf), axis=0, ignore_index=True)
with open('results/geojsondf.pkl', 'wb') as f:
pickle.dump(alldf, f)
return alldf
def merge_map_data(path, featuredf, store=False):
files = os.listdir(path)
files = [f[2:-4] for f in files if f[:2] == 'CL']
files.remove('xx')
# incomplete cut list
# files.remove('000104')
mapnos = [f for f in files if len(f) <= 6]
mapnos = ['020408']
fnums = [mapno2list(f) for f in mapnos]
# column names
fnames = map(lambda x: [FDICT[n] for n in x], fnums)
nclustersmax = 28
### make null map
cnum = cut2cluster('xx', nclustersmax, allowed_nodes=featuredf.index)
# retain only mutual nodes
nodelist = set(featuredf.index).intersection(set(cnum.index))
featuredf = featuredf.ix[nodelist]
cnum = cnum.ix[nodelist]
nclusters = len(cnum.unique())
clist = gencolors(nclusters)
rgmatrix = rg_colormatrix(most_similar(featuredf, cnum))
fbars = feature_bars(featuredf[FDICT.values()], cnum)
fn = 'data/uscensus/tl_2010_06075_tabblock10/tl_2010_06075_tabblock10.dbf'
mergedf = merge_shapefiles(featuredf[['lat', 'lon']], fn)
polys = make_shapefiles(featuredf[['lat', 'lon']], mergedf.polys, cnum)
# pdb.set_trace()
alldf = pd.DataFrame({'cnum': cnum.unique(),
'polygon': polys})
# alldf['color'] = clist
alldf['rgmatrix'] = map(lambda x: list(rgmatrix.ix[x]), cnum.unique())
alldf['mapno'] = ''
alldf['fbars'] = map(list, fbars.round(2).values)
# store results
if store:
alldf.to_csv('results/alldf.csv')
# make all other maps
for i, f in enumerate(mapnos):
print f
cnum = cut2cluster(f, nclustersmax, allowed_nodes=featuredf.index)
rgmatrix = rg_colormatrix(most_similar(featuredf, cnum))
# cnum = cnum.ix[nodelist]
fbars = feature_bars(featuredf[fnames[i]], cnum)
polys = make_shapefiles(featuredf[['lat', 'lon']],
mergedf.polys, cnum)
onedf = pd.DataFrame({'cnum': cnum.unique(),
'polygon': polys})
# onedf['color'] = clist
onedf['rgmatrix'] = map(lambda x: list(rgmatrix.ix[x]), cnum.unique())
onedf['mapno'] = f
onedf['fbars'] = map(list, fbars.round(2).values)
if store:
with open('results/alldf.csv', 'a') as storefile:
onedf.to_csv(storefile, header=False)
alldf = pd.concat((alldf, onedf), axis=0, ignore_index=True)
if store:
with open('results/alldf.pkl', 'wb') as f:
pickle.dump(alldf, f)
return alldf
