Merge pull request #94 from matsonj/add-nfl-model

Add-nfl-model

.github/workflows/deploy_on_netlify.yml

@@ -36,12 +36,11 @@ jobs:
       - name: evidence install & build
         run: |
           make evidence-build
-          cd analyze && npm run build
 
       - name: copy output to working directory
         run: |
           mkdir -p /home/runner/work/_temp/_github_home/build
-          cp -r analyze/build/* /home/runner/work/_temp/_github_home/build
+          cp -r evidence/build/* /home/runner/work/_temp/_github_home/build
 
       - name: Deploy Netlify
         uses: danakim/gh-action-deploy-netlify@master

.github/workflows/manual_build.yml

@@ -36,6 +36,7 @@ jobs:
           meltano --log-level=debug install loaders
           meltano --log-level=debug install mappers
           meltano --log-level=debug install utility dbt-duckdb
+          meltano --log-level=debug install utility evidence
         
 
       - name: build pipeline

.github/workflows/pr_to_master.yml

@@ -30,6 +30,7 @@ jobs:
           meltano --log-level=debug install loaders
           meltano --log-level=debug install mappers
           meltano --log-level=debug install utility dbt-duckdb
+          meltano --log-level=debug install utility evidence
 
       - name: build pipeline
         run: make pipeline

.gitignore

@@ -6,4 +6,4 @@
 /docs/run
 /data/data_catalog
 /output
-/logs
+/logs

Makefile

@@ -32,12 +32,19 @@ docker-run-superset:
 		mdsbox make pipeline superset-visuals
 
 evidence-build:
+	meltano invoke evidence upgrade
+	meltano invoke evidence build
+
+evidence-run:
+	meltano invoke evidence dev
+
+evidence-build-old:
 	cd analyze && npm i -force
 	cd analyze && mkdir -p data_catalog
 	cp -r data/data_catalog/* analyze/data_catalog
 	cp analyze/data_catalog/mdsbox.db analyze/
 
-evidence-run:
+evidence-run-old:
 	cd analyze && npm run dev -- --host 0.0.0.0
 
 evidence-visuals:

data/nfl/nfl_team_ratings.csv

@@ -0,0 +1,33 @@
+Team,Win Total,ELO rating,Conf,Division
+Buffalo Bills,10.5,1729.76563718597,AFC,East
+Miami Dolphins,9.5,1681.13241410486,AFC,East
+New York Jets,9.5,1676.81962490493,AFC,East
+New England Patriots,7.5,1605.49738694507,AFC,East
+Cincinnati Bengals,11.5,1717.78586912306,AFC,North
+Baltimore Ravens,8.5,1589.99596421381,AFC,North
+Cleveland Browns,8.5,1583.45094092808,AFC,North
+Pittsburgh Steelers,8.5,1578.09198940955,AFC,North
+Jacksonville Jaguars,9.5,1600.18698304192,AFC,South
+Tennessee Titans,7.5,1508.80023011146,AFC,South
+Indianapolis Colts,6.5,1459.76229611732,AFC,South
+Houston Texans,5.5,1418.28381122239,AFC,South
+Kansas City Chiefs,11.5,1762.52920677914,AFC,West
+Los Angeles Chargers,9.5,1669.57049071911,AFC,West
+Denver Broncos,8.5,1622.57658498554,AFC,West
+Las Vegas Raiders,7.5,1587.92769328906,AFC,West
+Philadelphia Eagles,10.5,1723.06689610956,NFC,East
+Dallas Cowboys,9.5,1680.12941645968,NFC,East
+New York Giants,8.5,1637.72667821245,NFC,East
+Washington Commanders,7.5,1596.99857761264,NFC,East
+Detroit Lions,9,1611.00993684181,NFC,North
+Minnesota Vikings,8.5,1607.42736606408,NFC,North
+Green Bay Packers,7.5,1549.21345906024,NFC,North
+Chicago Bears,7.5,1542.57634273391,NFC,North
+New Orleans Saints,9.5,1579.94137763577,NFC,South
+Carolina Panthers,7.5,1512.12835006181,NFC,South
+Atlanta Falcons,7.5,1501.03922942317,NFC,South
+Tampa Bay Buccaneers,6.5,1482.0264000382,NFC,South
+San Francisco 49ers,11.5,1730.12289472661,NFC,West
+Seattle Seahawks,8.5,1607.65690082201,NFC,West
+Los Angeles Rams,7.5,1566.85345285444,NFC,West
+Arizona Cardinals,5.5,1478.71549244838,NFC,West

docs/compiled/nba_monte_carlo/models/nba/conformed/inputs/latest_results.sql

@@ -0,0 +1,21 @@
+SELECT
+    (_smart_source_lineno - 1)::int AS game_id,
+    team1 AS home_team, 
+    score1 AS home_team_score,
+    team2 AS visiting_team,
+    score2 AS visiting_team_score,
+    date,
+    CASE 
+        WHEN score1 > score2 THEN team1
+        ELSE team2
+    END AS winning_team,
+    CASE 
+        WHEN score1 > score2 THEN team2
+        ELSE team1
+    END AS losing_team,
+    True AS include_actuals
+FROM "mdsbox"."main"."prep_nba_elo_latest"
+WHERE score1 IS NOT NULL 
+-- remove actual results so that we can sim 2023-24 season
+    AND 1 = 0
+GROUP BY ALL

docs/compiled/nba_monte_carlo/models/nba/conformed/inputs/post_season_schedule.sql

@@ -0,0 +1,13 @@
+SELECT
+    S.key::int AS game_id,
+    S.type,
+    S.series_id,
+    NULL AS visiting_conf,
+    S.visitorneutral AS visiting_team,
+    NULL AS visiting_team_elo_rating,
+    NULL AS home_conf,
+    S.homeneutral AS home_team,
+    NULL AS home_team_elo_rating
+FROM "mdsbox"."main"."prep_schedule" AS S
+WHERE S.type <> 'reg_season'
+GROUP BY ALL

docs/compiled/nba_monte_carlo/models/nba/conformed/inputs/random_num_gen.sql

@@ -0,0 +1,9 @@
+
+
+SELECT
+    i.scenario_id,
+    S.game_id,
+    (random() * 10000)::smallint AS rand_result,
+    0 AS sim_start_game_id
+FROM "mdsbox"."main"."scenario_gen" AS i
+CROSS JOIN "mdsbox"."main"."schedules" AS S

docs/compiled/nba_monte_carlo/models/nba/conformed/inputs/ratings.sql

@@ -0,0 +1,13 @@
+SELECT
+    orig.team,
+    orig.team_long,
+    orig.conf,
+    CASE
+        WHEN latest.latest_ratings = true AND latest.elo_rating IS NOT NULL THEN latest.elo_rating
+        ELSE orig.elo_rating
+    END AS elo_rating,
+    orig.elo_rating AS original_rating,
+    orig.win_total
+FROM "mdsbox"."main"."prep_team_ratings" orig
+LEFT JOIN "mdsbox"."main"."prep_elo_post" latest ON latest.team = orig.team
+GROUP BY ALL

docs/compiled/nba_monte_carlo/models/nba/conformed/inputs/reg_season_schedule.sql

@@ -0,0 +1,14 @@
+SELECT
+    (S._smart_source_lineno - 1)::int AS game_id,
+    'reg_season' AS type,
+    0 as series_id,
+    V.conf AS visiting_conf,
+    V.team AS visiting_team,
+    V.elo_rating::int AS visiting_team_elo_rating,
+    H.conf AS home_conf,
+    H.team AS home_team,
+    H.elo_rating::int AS home_team_elo_rating
+FROM "mdsbox"."main"."prep_nba_elo_latest" AS S
+LEFT JOIN "mdsbox"."main"."ratings" V ON V.team = S.team2
+LEFT JOIN "mdsbox"."main"."ratings" H ON H.team = S.team1
+GROUP BY ALL

docs/compiled/nba_monte_carlo/models/nba/conformed/inputs/scenario_gen.sql

@@ -0,0 +1,2 @@
+SELECT I.generate_series AS scenario_id
+FROM generate_series(1, 10000 ) AS I

docs/compiled/nba_monte_carlo/models/nba/conformed/inputs/schedules.sql

@@ -0,0 +1,7 @@
+SELECT
+    *
+FROM "mdsbox"."main"."reg_season_schedule"
+UNION ALL
+SELECT
+    *
+FROM "mdsbox"."main"."post_season_schedule"

docs/compiled/nba_monte_carlo/models/nba/conformed/inputs/teams.sql

@@ -0,0 +1,7 @@
+SELECT
+    S.visitorneutral AS team_long,
+    R.team
+FROM "mdsbox"."main"."prep_schedule" S
+LEFT JOIN "mdsbox"."main"."ratings" AS R ON R.team_long = S.visitorneutral
+WHERE R.team IS NOT NULL
+GROUP BY ALL

docs/compiled/nba_monte_carlo/models/nba/conformed/inputs/vegas_wins.sql

@@ -0,0 +1,5 @@
+SELECT
+    team,
+    win_total
+FROM "mdsbox"."main"."ratings"
+GROUP BY ALL

docs/compiled/nba_monte_carlo/models/nba/conformed/inputs/xf_series_to_seed.sql

@@ -0,0 +1,4 @@
+SELECT
+    series_id,
+    seed
+FROM "mdsbox"."main"."prep_xf_series_to_seed"

docs/compiled/nba_monte_carlo/models/nba/conformed/playin/playin_sim_r1.sql

@@ -0,0 +1,21 @@
+SELECT
+    R.scenario_id,
+    S.game_id,
+    EV.conf AS conf,
+    EV.winning_team AS visiting_team,
+    EV.elo_rating AS visiting_team_elo_rating,
+    EH.winning_team AS home_team,
+    EH.elo_rating AS home_team_elo_rating,
+    -- removing the 70 point adjustment for home team advantage for now
+   ( 1 - (1 / (10 ^ (-( EV.elo_rating - EH.elo_rating - 0)::real/400)+1))) * 10000 AS home_team_win_probability,
+    R.rand_result,
+    CASE 
+        WHEN -- removing the 70 point adjustment for home team advantage for now
+   ( 1 - (1 / (10 ^ (-( EV.elo_rating - EH.elo_rating - 0)::real/400)+1))) * 10000 >= R.rand_result THEN EH.winning_team
+        ELSE EV.winning_team
+    END AS winning_team 
+FROM "mdsbox"."main"."schedules" S
+    LEFT JOIN "mdsbox"."main"."random_num_gen" R ON R.game_id = S.game_id
+    LEFT JOIN "mdsbox"."main"."reg_season_end" EH ON S.home_team = EH.seed AND R.scenario_id = EH.scenario_id
+    LEFT JOIN "mdsbox"."main"."reg_season_end" EV ON S.visiting_team = EV.seed AND R.scenario_id = EV.scenario_id
+WHERE S.type = 'playin_r1'

docs/compiled/nba_monte_carlo/models/nba/conformed/playin/playin_sim_r1_end.sql

@@ -0,0 +1,32 @@
+WITH cte_playin_details AS (
+    SELECT
+        S.scenario_id,
+        S.game_id,
+        S.winning_team,
+        CASE
+            WHEN S.winning_team = S.home_team THEN S.home_team_elo_rating
+            ELSE S.visiting_team_elo_rating
+        END AS winning_team_elo_rating,
+        S.conf AS conf,
+        CASE
+            WHEN S.winning_team = S.home_team THEN S.visiting_team
+            ELSE S.home_team
+        END AS losing_team,
+        CASE
+            WHEN S.winning_team = S.home_team THEN S.visiting_team_elo_rating
+            ELSE S.home_team_elo_rating
+        END AS losing_team_elo_rating,
+        CASE
+            WHEN S.game_id IN (1231, 1234) THEN 'winner advance'
+            WHEN S.game_id IN (1232, 1235) THEN 'loser eliminated'
+        END AS result
+    FROM "mdsbox"."main"."playin_sim_r1" S
+)
+
+SELECT
+    *,
+    CASE
+        WHEN game_id IN (1231, 1234) THEN losing_team
+        WHEN game_id IN (1232, 1235) THEN winning_team
+    END AS remaining_team
+FROM cte_playin_details

docs/compiled/nba_monte_carlo/models/nba/conformed/playin/playin_sim_r2.sql

@@ -0,0 +1,23 @@
+SELECT 
+    R.scenario_id,
+    S.game_id,
+    S.home_team[7:] AS home_team_id,
+    S.visiting_team[8:] AS visiting_team_id,
+    EV.conf AS conf,
+    EV.remaining_team AS visiting_team,
+    EV.winning_team_elo_rating AS visiting_team_elo_rating,
+    EH.remaining_team AS home_team,
+    EH.losing_team_elo_rating AS home_team_elo_rating,
+    -- removing the 70 point adjustment for home team advantage for now
+   ( 1 - (1 / (10 ^ (-( EV.winning_team_elo_rating - EH.losing_team_elo_rating - 0)::real/400)+1))) * 10000 as home_team_win_probability,
+    R.rand_result,
+    CASE 
+        WHEN -- removing the 70 point adjustment for home team advantage for now
+   ( 1 - (1 / (10 ^ (-( EV.winning_team_elo_rating - EH.losing_team_elo_rating - 0)::real/400)+1))) * 10000 >= R.rand_result THEN EH.remaining_team
+        ELSE EV.remaining_team
+    END AS winning_team 
+FROM "mdsbox"."main"."schedules" S
+    LEFT JOIN "mdsbox"."main"."random_num_gen" R ON R.game_id = S.game_id
+    LEFT JOIN "mdsbox"."main"."playin_sim_r1_end" EH ON R.scenario_id = EH.scenario_id AND EH.game_id = S.home_team[7:]
+    LEFT JOIN "mdsbox"."main"."playin_sim_r1_end" EV ON R.scenario_id = EV.scenario_id AND EV.game_id = S.visiting_team[8:]
+WHERE S.type = 'playin_r2'

docs/compiled/nba_monte_carlo/models/nba/conformed/playin/playin_sim_r2_end.sql

@@ -0,0 +1,19 @@
+SELECT
+    P1.scenario_id,
+    P1.conf,
+    P1.winning_team,
+    P1.conf || '-7' AS seed,
+    P1.winning_team_elo_rating
+FROM "mdsbox"."main"."playin_sim_r1_end" P1
+WHERE P1.result = 'winner advance'
+UNION ALL
+SELECT
+    P2.scenario_id,
+    P2.conf AS conf,
+    P2.winning_team,
+    P2.conf || '-8' AS seed,
+    CASE
+        WHEN P2.winning_team = P2.home_team THEN P2.home_team_elo_rating
+        ELSE P2.visiting_team_elo_rating
+    END AS elo_rating
+FROM "mdsbox"."main"."playin_sim_r2" P2

docs/compiled/nba_monte_carlo/models/nba/conformed/playoffs/initialize_seeding.sql

@@ -0,0 +1,20 @@
+
+
+WITH cte_teams AS (
+    SELECT
+        scenario_id,
+        conf,
+        winning_team,
+        seed,
+        elo_rating
+    FROM "mdsbox"."main"."reg_season_end"
+    WHERE season_rank < 7
+    UNION ALL
+    SELECT *
+    FROM "mdsbox"."main"."playin_sim_r2_end"
+)
+
+SELECT
+    T.*,
+    0 AS sim_start_game_id
+FROM cte_teams T