databyjp

thresh_days = 20
df = df.assign(legend="Other teams")
df.loc[df["TEAM_NAME"] == highlight_tm, "legend"] = f"{thresh_days}+ days ago"
df.loc[(df.days_since_game < thresh_days) & (df["TEAM_NAME"] == highlight_tm), "legend"] = f"In the last {thresh_days} days"

fig = px.scatter(df, x="DEF_RATING", y="OFF_RATING",
                 color="legend",
                 title=f"Game-by-game performances by the {highlight_tm} - '21-'22",
                 color_discrete_sequence=["#dddddd", color_dict[highlight_tm][0], color_dict[highlight_tm][1]],
                 size="inv_days_since_game",

databyjp

import logging
import pandas as pd
import plotly.express as px
import json
import os
import utils

logger = logging.getLogger(__name__)
root_logger = logging.getLogger()
root_logger.setLevel(logging.INFO)

databyjp

def update_box_scores(box_json_dir="dl_data/box_scores/json"):
    """
    Download all box scores from NBA API
    :param box_json_dir: Directory to save downloads to
    :return:
    """
    logger.info("Starting download of game box scores...")
    gldf = utils.load_gamelogs()
    gldf = gldf.sort_values("gamedate_dt")
    gm_ids = gldf["Game_ID"].unique()

databyjp

def fetch_box_score(json_dir, gm_id):
    """
    Download one box score from NBA API
    :param json_dir: Directory for saving downloaded JSON
    :param gm_id: NBA game ID
    :return:
    """
    from nba_api.stats.endpoints import boxscoreadvancedv2
    dl_path = os.path.join(json_dir, gm_id + ".json")
    if os.path.exists(dl_path):

databyjp

       GAME_ID     TEAM_ID TEAM_ABBREVIATION      TEAM_CITY  PLAYER_ID         PLAYER_NAME  NICKNAME START_POSITION                 COMMENT    MIN  ...  TM_TOV_PCT  EFG_PCT  TS_PCT  USG_PCT  E_USG_PCT  E_PACE    PACE  PACE_PER40  POSS    PIE
0   0022100860  1610612760               OKC  Oklahoma City    1629647       Darius Bazley    Darius              F                          39:34  ...         0.0    0.529   0.573    0.213      0.215  105.83  104.33       86.94    85  0.115
1   0022100860  1610612760               OKC  Oklahoma City    1630598       Aaron Wiggins     Aaron              F                          33:50  ...         0.0    0.200   0.200    0.059      0.064  103.94  107.12       89.27    76  0.025
2   0022100860  1610612760               OKC  Oklahoma City    1629676         Isaiah Roby    Isaiah              C                          22:03  ...         0.0    0.500   0.436    0.132      0.132  101.01   97.96       81.63    46  0.020
3   0022100860  1610612760               OKC  Oklahoma C

databyjp

      GAME_ID     TEAM_ID TEAM_NAME TEAM_ABBREVIATION      TEAM_CITY     MIN  E_OFF_RATING  OFF_RATING  E_DEF_RATING  DEF_RATING  ...  TM_TOV_PCT  EFG_PCT  TS_PCT  USG_PCT  E_USG_PCT  E_PACE   PACE  PACE_PER40  POSS    PIE
0  0022100860  1610612760   Thunder               OKC  Oklahoma City  265:00         114.5       116.5         109.2       110.8  ...        10.1    0.568   0.589      1.0      0.202  101.23  99.62       83.02   109  0.531
1  0022100860  1610612752    Knicks               NYK       New York  265:00         109.2       110.8         114.5       116.5  ...        12.6    0.486   0.523      1.0      0.197  101.23  99.62       83.02   111  0.469
[2 rows x 29 columns]

databyjp

        temp_syms = [s for s in list(sp500_df[sp500_df.Sector == sect].Symbol.values) if s in df.symbol.unique()]
        temp_syms = random.sample(temp_syms, n_symbols)
        pf1_df = df[df.symbol.isin(temp_syms)]
        pf1_avg = get_avg_prices(pf1_df)
        pf1_var = pf1_avg.var()

databyjp

    # ========== DETERMINE SIMILARITIES ==========
    # Calculate similarities between each stock
    r_array = np.zeros([len(symbols), len(symbols)])
    p_array = np.zeros([len(symbols), len(symbols)])
    for i in range(len(symbols)):
        for j in range(len(symbols)):
            vals_i = df[df["symbol"] == symbols[i]]['close'].values
            vals_j = df[df["symbol"] == symbols[j]]['close'].values
            r_ij, p_ij = scipy.stats.pearsonr(vals_i, vals_j)
            r_array[i, j] = r_ij

databyjp

def normalise_price(df):
    df = df.assign(norm_close=0)
    for symbol in df["symbol"].unique():
        symbol_df = df[df["symbol"] == symbol]
        min_date = symbol_df["date"].min()
        ref_val = symbol_df[symbol_df["date"] == min_date]["close"].values[0]
        df.loc[df["symbol"] == symbol, "norm_close"] = df.loc[df["symbol"] == symbol, "close"] / ref_val
    return df

databyjp

    # ========== Get token ==========
    with open("../../tokens/iex_token.txt", "r") as f:
        iex_tkn = f.read().strip()

    # ========== Get symbols to use ==========
    symbol_list = ["MSFT", "AAPL", "NVDA", "JNJ", "KHC", "ALL"]

    # ========== Get symbols to use ==========
    date_range = '3m'
    symbol_dict = dict()