Understanding how dims='team' is connected to dataframe in Rugby example

Questions
1

Hello,

Trying to understand dims and coords, I am currently considering the Rugby example here

The model used for the article is the following:

with pm.Model(coords=coords) as model:
    # constant data
    home_team = pm.ConstantData("home_team", home_idx, dims="match")
    away_team = pm.ConstantData("away_team", away_idx, dims="match")

    # global model parameters
    home = pm.Normal("home", mu=0, sigma=1)
    sd_att = pm.HalfNormal("sd_att", sigma=2)
    sd_def = pm.HalfNormal("sd_def", sigma=2)
    intercept = pm.Normal("intercept", mu=3, sigma=1)

    # team-specific model parameters
    atts_star = pm.Normal("atts_star", mu=0, sigma=sd_att, dims="team")
    defs_star = pm.Normal("defs_star", mu=0, sigma=sd_def, dims="team")

    atts = pm.Deterministic("atts", atts_star - pt.mean(atts_star), dims="team")
    defs = pm.Deterministic("defs", defs_star - pt.mean(defs_star), dims="team")
    home_theta = pt.exp(intercept + home + atts[home_idx] + defs[away_idx])
    away_theta = pt.exp(intercept + atts[away_idx] + defs[home_idx])

    # likelihood of observed data
    home_points = pm.Poisson(
        "home_points",
        mu=home_theta,
        observed=df_all["home_score"],
        dims=("match"),
    )
    away_points = pm.Poisson(
        "away_points",
        mu=away_theta,
        observed=df_all["away_score"],
        dims=("match"),
    )
    trace = pm.sample(1000, tune=1500, cores=4)

I can see that this model uses ‘match’ and ‘team’ dimensions.

However, ‘match’ and ‘team’ are not used in a dataframe (I would expect them to be columns).

Screenshot 2023-04-12 at 22.15.01
Screenshot 2023-04-12 at 22.15.011078×398 29.6 KB

Can you please elaborate how ‘match’ and ‘team’ dimensions are connected to the dataset during the sampling? Should they be?

Thank you in advance!

1 Like
2

The “team” dimension and associated coordinates are created in the cell above:

home_idx, teams = pd.factorize(df_all["home_team"], sort=True)
away_idx, _ = pd.factorize(df_all["away_team"], sort=True)
coords = {"team": teams}

The “match” dimension is specified with the data:

home_team = pm.ConstantData("home_team", home_idx, dims="match")
away_team = pm.ConstantData("away_team", away_idx, dims="match")

This just names the dimension “match” to indicate that each value in home_team and each value in away_team is taken from a different match.

1 Like
3

Hello, @cluhmann!
Thanks a lot for your response, it’s clear now. Please excuse my late reply =)