diff --git a/gtfs_digest/deploy_portfolio_yaml.py b/gtfs_digest/deploy_portfolio_yaml.py
index 6537b2070..88db032b0 100644
--- a/gtfs_digest/deploy_portfolio_yaml.py
+++ b/gtfs_digest/deploy_portfolio_yaml.py
@@ -6,7 +6,7 @@
from segment_speed_utils.project_vars import RT_SCHED_GCS
-PORTFOLIO_SITE_YAML = Path("../portfolio/sites/gtfs_digest.yml")
+PORTFOLIO_SITE_YAML = Path("../portfolio/sites/digest_typologies.yml")
def overwrite_yaml(portfolio_site_yaml: Path) -> list:
"""
@@ -29,7 +29,19 @@ def overwrite_yaml(portfolio_site_yaml: Path) -> list:
districts = sorted(list(df.caltrans_district.unique()))
- operators = df.organization_name.tolist()
+ operators = df.organization_name.tolist()
+
+ keep_me = ["City of Santa Monica", "City of Culver City",
+ "Long Beach Transit", "Southern California Regional Rail Authority",
+ "Foothill Transit",
+ "Alameda-Contra Costa Transit District",
+ "City and County of San Francisco",
+ "Sonoma-Marin Area Rail Transit District",
+ "Marin County Transit District",
+ ]
+ operators = [i for i in operators if i in keep_me]
+ districts = sorted(df[df.organization_name.isin(keep_me)].caltrans_district.unique())
+
# Eric's example
# https://github.com/cal-itp/data-analyses/blob/main/rt_delay/04_generate_all.ipynb
diff --git a/gtfs_digest/extra_cleaning.py b/gtfs_digest/extra_cleaning.py
deleted file mode 100644
index b27893639..000000000
--- a/gtfs_digest/extra_cleaning.py
+++ /dev/null
@@ -1,28 +0,0 @@
-# operators that need route names parsed
-# probably just keep the long name portion
-# also do titlecase? this is finicky, because some are CC (which we don't want titlecase)
-operators_only_route_long_name = [
- "Antelope Valley Transit Authority Schedule",
- "Bay Area 511 ACE Schedule",
- "Bay Area 511 Caltrain Schedule",
- "Bay Area 511 Emery Go-Round Schedule",
- "Bay Area 511 Petaluma Schedule",
- "Beach Cities GMV Schedule",
- "Bear Schedule",
- "Commerce Schedule",
- "Elk Grove Schedule",
- "Humboldt Schedule",
- "LA DOT Schedule",
- "Lawndale Beat GMV Schedule",
- "Redding Schedule",
- "Redwood Coast Schedule",
- "Santa Maria Schedule",
- "StanRTA Schedule",
- "VCTC GMV Schedule",
- "Victor Valley GMV Schedule",
- "Visalia Schedule",
- "Yolobus Schedule",
-]
-
-# BruinBus Schedule - nothing shows up in route stats
-# Why does "StanRTA Schedule" and "Tahoe Transportation District Schedule" appear to have similar route names?
\ No newline at end of file
diff --git a/gtfs_digest/merge_data.py b/gtfs_digest/merge_data.py
index 3840ca992..7da589ce9 100644
--- a/gtfs_digest/merge_data.py
+++ b/gtfs_digest/merge_data.py
@@ -136,54 +136,37 @@ def concatenate_crosswalk_organization(
def merge_in_standardized_route_names(
df: pd.DataFrame,
) -> pd.DataFrame:
+
keep_cols = [
"schedule_gtfs_dataset_key", "name",
"route_id", "service_date",
- "recent_route_id2", "recent_combined_name"]
+ ]
CLEAN_ROUTES = GTFS_DATA_DICT.schedule_tables.route_identification
- operators_need_cleaning = pd.read_parquet(
- f"{SCHED_GCS}{CLEAN_ROUTES}.parquet",
- filters = [[("name", "in", operators_only_route_long_name)]]
- )
- operators_ok = pd.read_parquet(
- f"{SCHED_GCS}{CLEAN_ROUTES}.parquet",
- filters = [[("name", "not in", operators_only_route_long_name)]]
- )
+ route_names_df = pd.read_parquet(f"{SCHED_GCS}{CLEAN_ROUTES}.parquet")
- operators_need_cleaning = operators_need_cleaning.assign(
- recent_combined_name = operators_need_cleaning.route_long_name
- )
+ route_names_df = time_series_utils.clean_standardized_route_names(
+ route_names_df).drop_duplicates()
- standardized_route_names = pd.concat([
- operators_need_cleaning,
- operators_ok
- ], axis=0, ignore_index=True)[keep_cols]
-
if "name" in df.columns:
df = df.drop(columns = "name")
- df = pd.merge(
+ # Use `route_id` to merge to standardized_route_names
+ df2 = pd.merge(
df,
- standardized_route_names,
+ route_names_df,
on = ["schedule_gtfs_dataset_key",
"route_id", "service_date"],
how = "left",
)
- df = df.assign(
- recent_combined_name = df.recent_combined_name.str.replace("__", " ")
- ).drop(
- columns = ["route_id"]
- ).rename(
- columns = {
- "recent_route_id2": "route_id",
- "recent_combined_name": "route_combined_name"
- }
- )
+ # After merging, we can replace route_id with recent_route_id2
+ drop_cols = ["route_desc", "combined_name", "route_id2"]
+ df3 = time_series_utils.parse_route_combined_name(df2).drop(
+ columns = drop_cols)
- return df
+ return df3
if __name__ == "__main__":
@@ -237,7 +220,7 @@ def merge_in_standardized_route_names(
"is_early", "is_ontime", "is_late"
]
- df[integrify] = df[integrify].astype("Int64")
+ df[integrify] = df[integrify].fillna(0).astype("int")
df.to_parquet(
f"{RT_SCHED_GCS}{DIGEST_RT_SCHED}.parquet"
diff --git a/gtfs_digest/readable.yml b/gtfs_digest/readable.yml
new file mode 100644
index 000000000..6e2982733
--- /dev/null
+++ b/gtfs_digest/readable.yml
@@ -0,0 +1,66 @@
+# GTFS schedule
+direction_id:
+ readable: Direction
+ caption: Something
+avg_scheduled_service_minutes: "Average Scheduled Service (trip minutes)"
+avg_stop_miles: Average Stop Distance (miles)
+ttl_service_hours: total service (hours)
+n_scheduled_trips: "# scheduled trips"
+frequency: Trips per Hour
+total_scheduled_service_minutes: "Aggregate Scheduled Service Minutes (all trips)"
+route_id: Route
+route_combined_name: Route
+
+# GTFS vehicle positions
+minutes_atleast1_vp: "# minutes with 1+ vp per minute"
+minutes_atleast2_vp: "# minutes with 2+ vp per minute"
+total_rt_service_minutes: "Aggregate Actual Service Minutes (all trips)"
+total_vp: "# vp"
+vp_in_shape: "# vp within scheduled shape"
+is_early": "# early arrival trips"
+is_ontime: "# on-time trips"
+is_late: "# late trips"
+n_vp_trips: "# trips with vp"
+vp_per_minute: "Average vp per minute"
+pct_in_shape: "% vp within scheduled shape"
+pct_rt_journey_atleast1_vp: "% actual trip minutes with 1+ vp per minute"
+pct_rt_journey_atleast2_vp: "% actual trip minutes with 2+ vp per minute"
+pct_sched_journey_atleast1_vp: "% scheduled trip minutes with 1+ vp per minute"
+pct_sched_journey_atleast2_vp: "% scheduled trip minutes with 2+ vp per minute"
+rt_sched_journey_ratio: "Actual / scheduled service ratio"
+avg_rt_service_minutes: "Average Actual Service (trip minutes)"
+speed_mph: "Speed (mph)"
+sched_rt_category: "GTFS Availability"
+
+# Operators
+organization_source_record_id: "Organization ID"
+organization_name: Organization
+name: Transit Operator
+caltrans_district: District
+base64_url: "base64 encoded feed URL"
+operator_n_routes: "# routes"
+operator_n_trips: "# trips"
+operator_n_shapes: "# shapes"
+operator_n_stops: "# stops"
+operator_n_arrivals: "# arrivals"
+operator_route_length_miles: "service miles"
+operator_arrivals_per_stop: "avg arrivals per stop"
+n_coverage_routes: "# coverage route types"
+n_downtown_local_routes: "# downtown local route types"
+n_local_routes: "# local route types"
+n_rapid_routes: "# rapid route types"
+
+
+# Dates / time
+time_period: Period
+service_date: Date
+time_of_day: "time of day"
+day_name: "Day of Week"
+year: Year
+month: Month
+year_month: "Month - Year"
+
+# Roads
+road_freq_category: "Road Frequency Category"
+road_typology: "Road Typology Category"
+pct_typology: "% route miles with typology"
\ No newline at end of file
diff --git a/gtfs_funnel/concatenate_monthly_scheduled_service.py b/gtfs_funnel/concatenate_monthly_scheduled_service.py
index d49f70db9..85596e3fb 100644
--- a/gtfs_funnel/concatenate_monthly_scheduled_service.py
+++ b/gtfs_funnel/concatenate_monthly_scheduled_service.py
@@ -14,7 +14,8 @@
MONTHLY_SERVICE = GTFS_DATA_DICT.schedule_tables.monthly_scheduled_service
CROSSWALK = GTFS_DATA_DICT.schedule_tables.gtfs_key_crosswalk
-
+ ROUTES = GTFS_DATA_DICT.schedule_tables.route_identification
+
year_list = [2023, 2024]
analysis_date_list = (rt_dates.y2024_dates +
rt_dates.y2023_dates +
@@ -47,11 +48,24 @@
"schedule_source_record_id",
"schedule_gtfs_dataset_key",
"organization_source_record_id", "organization_name",
- ],
+ ]
).drop(
columns = "service_date"
- ).drop_duplicates().reset_index(drop=True)
+ ).drop_duplicates().reset_index(drop=True)
+
+ # Get standardized route names and clean up more
+ standardized_routes = pd.read_parquet(f"{SCHED_GCS}{ROUTES}.parquet")
+
+ route_names_df = time_series_utils.clean_standardized_route_names(
+ standardized_routes).pipe(
+ time_series_utils.parse_route_combined_name
+ )[
+ ["schedule_gtfs_dataset_key",
+ "route_long_name", "route_short_name",
+ "route_id", "route_combined_name"]
+ ].drop_duplicates()
+ # Merge monthly service with crosswalk to get schedule_gtfs_dataset_key
df2 = pd.merge(
df,
crosswalk,
@@ -59,6 +73,15 @@
how = "inner",
)
- df2.to_parquet(
+ # Merge in route_names so we use the standardized/cleaned up route names
+ df3 = pd.merge(
+ df2,
+ route_names_df,
+ on = ["schedule_gtfs_dataset_key",
+ "route_long_name", "route_short_name"],
+ how = "left",
+ )
+
+ df3.to_parquet(
f"{SCHED_GCS}{MONTHLY_SERVICE}.parquet"
)
\ No newline at end of file
diff --git a/portfolio/digest_typologies/README.md b/portfolio/digest_typologies/README.md
new file mode 100644
index 000000000..4aca9d8a1
--- /dev/null
+++ b/portfolio/digest_typologies/README.md
@@ -0,0 +1,5 @@
+# GTFS Digest
+
+Performance metrics from GTFS schedule and vehicle positions time-series data for all transit operators by route.
+
+[](https://mermaid.live/edit#pako:eNqVk1Fv2jAQx7-K5QnlJUGEhAKeNKkU2NOeWu1hZIpMfAGrjh3Zl7WA-O5zknXQba20PDin-_98Z9-dT7QwAiijURRlGiUqYOTzw_qeLOUOHGa6EwaDU6YJkVoiI51JSIB7qCBgJNhyB0F47f3KreRbBS74jXuptrLi9nBnlLHtvg-ryXqyvn3ZeiEe4Bkv1Gg0-htZGCvAvgUpqeEtzUFhtHh9jvV6ulpcMQgW5SukLMugl8_tzy_nwSDTmS6VeSr23CJ5WPSAa7Y7y-s9saZBiIS0UKA0ulf7tVDcuSWURJSklEqxX7X4Q3WFlTW-EN0pr6O03-3GFXsQjQJSAVpZuO-MMR82ij4R0TXx44VebFxTtfUjrgYQ_2CvQt9tLOY_XP4_CZabHZYuf4RDjiY3dse1PPL2-qSwxrknrh7f27_aOORte4Q8gsi7EuaaV_Bu1t4CLS4uGtIKbMWl8NPdzWBGu9nMKPOmgJI3CjPqG-lR3qC5P-iCMrQNhLSpBUdYSu47WVFWcuW8F4REY7_0L6Z7OCGtuabsRJ8pi2bjYTJPxvM0mc6S-ThOQ3qgLLkZxulklt6kaTwZzdL4HNKjMT5qPBzHfjZH8TRJPJ1O0y7ct07sUp5_AvaOGGY)
\ No newline at end of file
diff --git a/portfolio/digest_typologies/_config.yml b/portfolio/digest_typologies/_config.yml
new file mode 100644
index 000000000..89e63588f
--- /dev/null
+++ b/portfolio/digest_typologies/_config.yml
@@ -0,0 +1,43 @@
+# Book settings
+# Learn more at https://jupyterbook.org/customize/config.html
+
+title: Transit Operators and Route Typologies
+author: Cal-ITP
+copyright: "2024"
+#logo: calitp_logo_MAIN.png
+
+# Force re-execution of notebooks on each build.
+# See https://jupyterbook.org/content/execute.html
+execute:
+ execute_notebooks: 'off'
+ allow_errors: false
+ timeout: -1
+
+# Define the name of the latex output file for PDF builds
+latex:
+ latex_documents:
+ targetname: book.tex
+
+launch_buttons:
+ binderhub_url: "https://mybinder.org"
+ jupyterhub_url: "https://hubtest.k8s.calitp.jarv.us"
+ thebe: true
+
+repository:
+ url: https://github.com/cal-itp/data-analyses/ # Online location of your book
+# path_to_book: docs # Optional path to your book, relative to the repository root
+ path_to_book: gtfs_digest
+ branch: main # Which branch of the repository should be used when creating links (optional)
+
+# Add GitHub buttons to your book
+# See https://jupyterbook.org/customize/config.html#add-a-link-to-your-repository
+html:
+ use_issues_button: true
+ use_repository_button: true
+ use_edit_page_button: true
+ google_analytics_id: 'G-JCX3Z8JZJC'
+
+sphinx:
+ config:
+ html_js_files:
+ - https://cdnjs.cloudflare.com/ajax/libs/require.js/2.3.4/require.min.js
\ No newline at end of file
diff --git a/portfolio/digest_typologies/_toc.yml b/portfolio/digest_typologies/_toc.yml
new file mode 100644
index 000000000..7ec7bfc5c
--- /dev/null
+++ b/portfolio/digest_typologies/_toc.yml
@@ -0,0 +1,11 @@
+format: jb-book
+parts:
+- caption: null
+ chapters:
+ - file: district_04-oakland.md
+ sections:
+ - glob: district_04-oakland/*
+ - file: district_07-los-angeles.md
+ sections:
+ - glob: district_07-los-angeles/*
+root: README
diff --git a/portfolio/digest_typologies/district_04-oakland.md b/portfolio/digest_typologies/district_04-oakland.md
new file mode 100644
index 000000000..1faedbdab
--- /dev/null
+++ b/portfolio/digest_typologies/district_04-oakland.md
@@ -0,0 +1 @@
+# District 04 - Oakland
\ No newline at end of file
diff --git a/portfolio/digest_typologies/district_04-oakland/0__typologies__district_04-oakland__name_alameda-contra-costa-transit-district.ipynb b/portfolio/digest_typologies/district_04-oakland/0__typologies__district_04-oakland__name_alameda-contra-costa-transit-district.ipynb
new file mode 100644
index 000000000..c2a2bd74e
--- /dev/null
+++ b/portfolio/digest_typologies/district_04-oakland/0__typologies__district_04-oakland__name_alameda-contra-costa-transit-district.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1276c09765845621b2dc6bb6675d61d097fd8a6ec6fb14da638a47026bab5a4
+size 12303490
diff --git a/portfolio/digest_typologies/district_04-oakland/1__typologies__district_04-oakland__name_central-contra-costa-transit-authority.ipynb b/portfolio/digest_typologies/district_04-oakland/1__typologies__district_04-oakland__name_central-contra-costa-transit-authority.ipynb
new file mode 100644
index 000000000..ec6ba3b25
--- /dev/null
+++ b/portfolio/digest_typologies/district_04-oakland/1__typologies__district_04-oakland__name_central-contra-costa-transit-authority.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:506dc5a99b5aa462e3b4e3ab2ec9012d575c6417a05a38549e6b8f2b2f1066ee
+size 3976205
diff --git a/portfolio/digest_typologies/district_04-oakland/2__typologies__district_04-oakland__name_city-and-county-of-san-francisco.ipynb b/portfolio/digest_typologies/district_04-oakland/2__typologies__district_04-oakland__name_city-and-county-of-san-francisco.ipynb
new file mode 100644
index 000000000..5d070b04e
--- /dev/null
+++ b/portfolio/digest_typologies/district_04-oakland/2__typologies__district_04-oakland__name_city-and-county-of-san-francisco.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:606bbf751b2557e15390740232a686baf88106fe0a8f9ea38f6ba4967ada83cb
+size 3285371
diff --git a/portfolio/digest_typologies/district_04-oakland/3__typologies__district_04-oakland__name_eastern-contra-costa-transit-authority.ipynb b/portfolio/digest_typologies/district_04-oakland/3__typologies__district_04-oakland__name_eastern-contra-costa-transit-authority.ipynb
new file mode 100644
index 000000000..677c41fcb
--- /dev/null
+++ b/portfolio/digest_typologies/district_04-oakland/3__typologies__district_04-oakland__name_eastern-contra-costa-transit-authority.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b9247a3914d486cba04475280240c715bbb9dd77c4296581586b65c1131bfd9
+size 1193675
diff --git a/portfolio/digest_typologies/district_04-oakland/4__typologies__district_04-oakland__name_marin-county-transit-district.ipynb b/portfolio/digest_typologies/district_04-oakland/4__typologies__district_04-oakland__name_marin-county-transit-district.ipynb
new file mode 100644
index 000000000..ceaae43e2
--- /dev/null
+++ b/portfolio/digest_typologies/district_04-oakland/4__typologies__district_04-oakland__name_marin-county-transit-district.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7156f50617a1067405dd4ec39a396a6948415b26f89b80f14fba2ee5231a2d5
+size 1926753
diff --git a/portfolio/digest_typologies/district_04-oakland/5__typologies__district_04-oakland__name_santa-clara-valley-transportation-authority.ipynb b/portfolio/digest_typologies/district_04-oakland/5__typologies__district_04-oakland__name_santa-clara-valley-transportation-authority.ipynb
new file mode 100644
index 000000000..b0affecd0
--- /dev/null
+++ b/portfolio/digest_typologies/district_04-oakland/5__typologies__district_04-oakland__name_santa-clara-valley-transportation-authority.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b3b8b1172c30472762b20fe87bbb16653c30e7f21c38114606af41b087c97e2c
+size 4497864
diff --git a/portfolio/digest_typologies/district_04-oakland/6__typologies__district_04-oakland__name_sonoma-county.ipynb b/portfolio/digest_typologies/district_04-oakland/6__typologies__district_04-oakland__name_sonoma-county.ipynb
new file mode 100644
index 000000000..32fe043e6
--- /dev/null
+++ b/portfolio/digest_typologies/district_04-oakland/6__typologies__district_04-oakland__name_sonoma-county.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:05d2cdf6ca57f45fb00c5dedbce8287fba685f851e14157d7eb024a15792d5a0
+size 928009
diff --git a/portfolio/digest_typologies/district_07-los-angeles.md b/portfolio/digest_typologies/district_07-los-angeles.md
new file mode 100644
index 000000000..79901aa5a
--- /dev/null
+++ b/portfolio/digest_typologies/district_07-los-angeles.md
@@ -0,0 +1 @@
+# District 07 - Los Angeles
\ No newline at end of file
diff --git a/portfolio/digest_typologies/district_07-los-angeles/0__typologies__district_07-los-angeles__name_city-of-culver-city.ipynb b/portfolio/digest_typologies/district_07-los-angeles/0__typologies__district_07-los-angeles__name_city-of-culver-city.ipynb
new file mode 100644
index 000000000..b54bfae5e
--- /dev/null
+++ b/portfolio/digest_typologies/district_07-los-angeles/0__typologies__district_07-los-angeles__name_city-of-culver-city.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f53075c3f2f59b845694cf8b0074e4b51210b0175f7aa9f3102812f9e1bddcdd
+size 917899
diff --git a/portfolio/digest_typologies/district_07-los-angeles/1__typologies__district_07-los-angeles__name_city-of-los-angeles.ipynb b/portfolio/digest_typologies/district_07-los-angeles/1__typologies__district_07-los-angeles__name_city-of-los-angeles.ipynb
new file mode 100644
index 000000000..ab90ab766
--- /dev/null
+++ b/portfolio/digest_typologies/district_07-los-angeles/1__typologies__district_07-los-angeles__name_city-of-los-angeles.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5ce36d0c5af0424b262b29cc0e3186d0938ef9339c1dc83455b0df988f3355e
+size 2841414
diff --git a/portfolio/digest_typologies/district_07-los-angeles/2__typologies__district_07-los-angeles__name_city-of-santa-monica.ipynb b/portfolio/digest_typologies/district_07-los-angeles/2__typologies__district_07-los-angeles__name_city-of-santa-monica.ipynb
new file mode 100644
index 000000000..adbcd9571
--- /dev/null
+++ b/portfolio/digest_typologies/district_07-los-angeles/2__typologies__district_07-los-angeles__name_city-of-santa-monica.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3f441f04cbd5695e09431c1236cb156cde07217b709ba0d7631fe5ddfce22df
+size 1119444
diff --git a/portfolio/digest_typologies/district_07-los-angeles/3__typologies__district_07-los-angeles__name_foothill-transit.ipynb b/portfolio/digest_typologies/district_07-los-angeles/3__typologies__district_07-los-angeles__name_foothill-transit.ipynb
new file mode 100644
index 000000000..90f0e7a58
--- /dev/null
+++ b/portfolio/digest_typologies/district_07-los-angeles/3__typologies__district_07-los-angeles__name_foothill-transit.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b432855e8750c47c8758130b947af4b78429d096da6928436299179f2b9fe1e
+size 1439613
diff --git a/portfolio/digest_typologies/district_07-los-angeles/4__typologies__district_07-los-angeles__name_long-beach-transit.ipynb b/portfolio/digest_typologies/district_07-los-angeles/4__typologies__district_07-los-angeles__name_long-beach-transit.ipynb
new file mode 100644
index 000000000..2b33c1fc3
--- /dev/null
+++ b/portfolio/digest_typologies/district_07-los-angeles/4__typologies__district_07-los-angeles__name_long-beach-transit.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b043b8a033cc5d219362d984759971bb6655f284def9435878f559b0d0c72c3f
+size 2798382
diff --git a/portfolio/digest_typologies/district_07-los-angeles/5__typologies__district_07-los-angeles__name_los-angeles-county-metropolitan-transportation-authority.ipynb b/portfolio/digest_typologies/district_07-los-angeles/5__typologies__district_07-los-angeles__name_los-angeles-county-metropolitan-transportation-authority.ipynb
new file mode 100644
index 000000000..2ea04f903
--- /dev/null
+++ b/portfolio/digest_typologies/district_07-los-angeles/5__typologies__district_07-los-angeles__name_los-angeles-county-metropolitan-transportation-authority.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c75d92ac12f4c899139b1c3146098e15f8ff0be9ea1ff160bb992cda77011d9f
+size 8359278
diff --git a/portfolio/digest_typologies/district_07-los-angeles/6__typologies__district_07-los-angeles__name_university-of-california-los-angeles.ipynb b/portfolio/digest_typologies/district_07-los-angeles/6__typologies__district_07-los-angeles__name_university-of-california-los-angeles.ipynb
new file mode 100644
index 000000000..6d2b8e7f9
--- /dev/null
+++ b/portfolio/digest_typologies/district_07-los-angeles/6__typologies__district_07-los-angeles__name_university-of-california-los-angeles.ipynb
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8536176aae1e3fd2c2528e0f412e358608338ce546c720df240a4a67bd5da311
+size 313182
diff --git a/portfolio/sites/digest_typologies.yml b/portfolio/sites/digest_typologies.yml
new file mode 100644
index 000000000..0baba7a4c
--- /dev/null
+++ b/portfolio/sites/digest_typologies.yml
@@ -0,0 +1,28 @@
+directory: ./gtfs_digest/
+notebook: ./gtfs_digest/typologies.ipynb
+parts:
+- chapters:
+ - caption: District 04 - Oakland
+ params:
+ district: 04 - Oakland
+ sections:
+ - name: Alameda-Contra Costa Transit District
+ - name: Central Contra Costa Transit Authority
+ - name: City and County of San Francisco
+ - name: Eastern Contra Costa Transit Authority
+ - name: Marin County Transit District
+ - name: Santa Clara Valley Transportation Authority
+ - name: Sonoma County
+ - caption: District 07 - Los Angeles
+ params:
+ district: 07 - Los Angeles
+ sections:
+ - name: City of Culver City
+ - name: City of Los Angeles
+ - name: City of Santa Monica
+ - name: Foothill Transit
+ - name: Long Beach Transit
+ - name: Los Angeles County Metropolitan Transportation Authority
+ - name: University of California, Los Angeles
+readme: ./gtfs_digest/README.md
+title: Transit Operators and Route Typologies
diff --git a/rt_segment_speeds/segment_speed_utils/time_series_utils.py b/rt_segment_speeds/segment_speed_utils/time_series_utils.py
index 949ca85ac..709fa9936 100644
--- a/rt_segment_speeds/segment_speed_utils/time_series_utils.py
+++ b/rt_segment_speeds/segment_speed_utils/time_series_utils.py
@@ -12,7 +12,7 @@
from typing import Literal
from segment_speed_utils import helpers
-from segment_speed_utils.project_vars import SEGMENT_GCS
+from segment_speed_utils.project_vars import SCHED_GCS, SEGMENT_GCS
fs = gcsfs.GCSFileSystem()
@@ -60,4 +60,68 @@ def concatenate_datasets_across_dates(
if get_pandas:
df = compute(df)[0]
- return df
\ No newline at end of file
+ return df
+
+
+def clean_standardized_route_names(
+ df: pd.DataFrame,
+) -> pd.DataFrame:
+ """
+ Clean up route names for operators that need
+ additional parsing. Just keep the route_long_name instead of
+ combining it with route_short_name.
+
+ TODO: do titlecase?
+ this is finicky, because some are CC (which we don't want titlecase)
+ """
+ df_need_cleaning = df.loc[df.name.isin(operators_only_route_long_name)]
+
+ df_ok = df.loc[~df.name.isin(operators_only_route_long_name)]
+
+ df_need_cleaning = df_need_cleaning.assign(
+ recent_combined_name = df_need_cleaning.route_long_name
+ )
+
+ df2 = pd.concat([
+ df_need_cleaning,
+ df_ok
+ ], axis=0, ignore_index=True)
+
+ return df2
+
+def parse_route_combined_name(df):
+ df = df.assign(
+ recent_combined_name = df.recent_combined_name.str.replace("__", " ")
+ ).drop(
+ columns = ["route_id"]
+ ).rename(
+ columns = {
+ "recent_route_id2": "route_id",
+ "recent_combined_name": "route_combined_name"
+ }
+ )
+
+ return df
+
+operators_only_route_long_name = [
+ "Antelope Valley Transit Authority Schedule",
+ "Bay Area 511 ACE Schedule",
+ "Bay Area 511 Caltrain Schedule",
+ "Bay Area 511 Emery Go-Round Schedule",
+ "Bay Area 511 Petaluma Schedule",
+ "Beach Cities GMV Schedule",
+ "Bear Schedule",
+ "Commerce Schedule",
+ "Elk Grove Schedule",
+ "Humboldt Schedule",
+ "LA DOT Schedule",
+ "Lawndale Beat GMV Schedule",
+ "Redding Schedule",
+ "Redwood Coast Schedule",
+ "Santa Maria Schedule",
+ "StanRTA Schedule",
+ "VCTC GMV Schedule",
+ "Victor Valley GMV Schedule",
+ "Visalia Schedule",
+ "Yolobus Schedule",
+]
\ No newline at end of file