def update_team_comparison(year, indicators, metric, hover_data, click_data):
df1 = subset_years(df_eoy, year)
dff = df1.melt(id_vars='team', value_vars=indicators, var_name='indicator')
dff['indicator'] = pd.Categorical(dff.indicator, indicators)
dff = pd.merge(dff, palette_df, how='outer')
dff = dff.sort_values(['indicator', 'team'], ascending=[False, True])
dff["rank"] = dff.groupby("indicator")["value"].rank("min",
ascending=False)
if hover_data:
i = hover_data['points'][0]['curveNumber']
elif click_data:
i = click_data['points'][0]['curveNumber']
else:
i = 0
marker_sizes = [10] * len(dff.team.unique())
line_widths = [0.4] * len(dff.team.unique())
marker_sizes[i] = 20
line_widths[i] = 2
metric2 = "value"
if metric == "value":
metric2 = "rank"
return {
'data': [
go.Scatter(
x=dff[dff.team == t][metric],
y=dff[dff.team == t]['indicator'],
name=t,
text=round(dff[dff.team == t][metric2], 2),
mode='markers+lines',
marker={
'size': marker_sizes[i], # 10
'color': dff[dff.team == t]['color1'],
'line': {
'width': 2,
'color': dff[dff.team == t]['color2']
}
},
line={
'width': line_widths[i],
'color': map_colors(t, palette, 1)
}) for i, t in enumerate(dff.team.unique())
],
'layout':
go.Layout(
# title=team,
xaxis=dict(title=metric, titlefont=dict(size=18)),
height=200 + 100 * len(indicators),
margin={
'l': 200,
'b': 40,
't': 40,
'r': 40
},
hovermode='closest')
}
def update_matchup_heatmap(division, year, indicator):
df1 = subset_years(df_t_wide, year)
teams = division_dict[division]
opponents = division_dict[division]
df1 = df1.loc[(df1.team.isin(teams)) & (df1.opponent.isin(opponents))]
df1 = df1.groupby(['team', 'opponent'])[indicator].mean().reset_index()
df1 = df1.sort_values(['team', 'opponent'], ascending=[False, True]) #
return {
'data': [
go.Heatmap(z=round(df1[indicator], 2),
x=df1.opponent,
y=df1.team,
reversescale=True,
text=[indicator] * len(df1),
hoverinfo='z+text')
],
'layout':
go.Layout(title=f'Average {indicator} by matchup',
margin={
'l': 200,
'b': 40,
't': 40,
'r': 40
},
yaxis=dict(title='Team'),
xaxis=dict(title='Opponent'))
}
def update_players(team, year, indicators, min_games, rate_type, hover_data,
click_data):
df1 = subset_years(df_p, year)
df1 = df1[df1.team == team]
df1 = apply_game_threshold(df1, n_games=min_games)
dff = df1.groupby('player')[player_indicators].sum().reset_index()
dff = aggregate_rates(dff, player_indicators, rate_type)
dff = dff.melt(id_vars='player',
value_vars=indicators,
var_name='indicator')
dff['indicator'] = pd.Categorical(dff.indicator, indicators)
dff = dff.sort_values(['indicator', 'player'], ascending=[False, True])
players = dff.player.unique()
if hover_data:
i = hover_data['points'][0]['curveNumber']
elif click_data:
i = click_data['points'][0]['curveNumber']
else:
i = 0
opacities = [0.5] * len(players)
marker_sizes = [10] * len(players)
line_widths = [0.4] * len(players)
opacities[i] = 1
marker_sizes[i] = 20
line_widths[i] = 2
return {
'data': [
go.Scatter(
x=dff[dff.player == p]['value'],
y=dff[dff.player == p]['indicator'],
name=p,
mode='markers+lines',
marker={
'size': marker_sizes[i],
'opacity': opacities[i],
# 'line': {'width': 0}
},
line={'width': line_widths[i]}) for i, p in enumerate(players)
],
'layout':
go.Layout(
# title=team,
height=80 + 80 * len(indicators),
margin={
'l': 120,
'b': 40,
't': 40,
'r': 40
},
hovermode='closest')
}
def make_conversion_plot(team, year, line):
"""TODO - refactor so easier to move to viz utils, for now it has too many dependencies"""
df1 = subset_years(df_p, year)
df1 = df1[df1.team == team]
# Todo - add this to player indicators?
df1['Proportion Offensive'] = df1[f'O Points Played'] / df1['Points Played']
title = ""
if line == 'O':
title = "Offense"
elif line == 'D':
title = "Defense"
return {
'data': [
go.Scatter(
x=df1[f'{line} Points Played'],
y=df1[f'{line}-line Scoring Efficiency'],
# name=p,
mode='markers',
marker=dict(
color=df1['Proportion Offensive'],
colorscale='RdBu',
reversescale=True,
showscale=True,
colorbar=dict(title='% Offensive', ),
),
text=df1['player'],
# 'size': 10,
# 'opacity': 0.5,
# 'line': {'width': 0}
# line={'width': 0.4}
)
],
'layout':
go.Layout(
# title=title,
height=400,
margin={
'l': 120,
'b': 40,
't': 40,
'r': 40
},
hovermode='closest',
yaxis=dict(
range=[0, 1],
title=f'{line}-line Scoring Efficiency',
),
xaxis=dict(title=f'Points Played ({line})'))
}
def update_leaderboard(indicator, year, rate_type, min_games):
df1 = subset_years(df_p, year)
df1 = apply_game_threshold(df1, n_games=min_games)
dff = df1.groupby('player')[player_indicators].sum().reset_index()
dff = aggregate_rates(dff, player_indicators, rate_type)
dff = dff.melt(id_vars='player',
value_vars=player_indicators,
var_name='indicator')
n_players = 20
dff = dff[dff.indicator == indicator]
# Get correct team & color for specific year
if year == 'All seasons':
player_map = player_team[player_team['year'] == 2018]
else:
player_map = player_team[player_team['year'] == year]
dff = pd.merge(player_map, dff)
dff = pd.merge(dff, palette_df,
how='outer').sort_values('value',
ascending=False)[:n_players]
dff = dff.sort_values('value',
ascending=True) # plotly seems to invert the order?
return {
'data': [
go.Scatter(
x=dff['value'],
y=dff['player'],
text=dff['team'],
mode='markers',
marker={
'size': 15,
'color': dff['color1'],
'line': {
'width': 3,
'color': dff['color2']
}
},
)
],
'layout':
go.Layout(
# title=indicator,
height=600,
margin={
'l': 120,
'b': 40,
't': 40,
'r': 0
},
hovermode='closest')
}
def update_team_timeseries(team, year, indicators):
df1 = subset_years(df_t, year)
df1 = df1[df1.team == team]
dff = df1.sort_values(['indicator', 'date'], ascending=[False, True])
return {
'data': [
go.Scatter(
x=dff[dff.indicator == i]['date'],
y=dff[dff.indicator == i]['value'],
name=i,
text=dff['opponent'],
mode='lines+markers',
marker={
'size': 15,
# 'opacity': 0.5,
'color': dff['opponent_color1'],
'line': {
'width': 3,
'color': dff['opponent_color2']
}
},
line={'width': 3}) for i in indicators
],
'layout':
go.Layout(
# title=team,
height=400,
margin={
'l': 120,
'b': 40,
't': 40,
'r': 40
},
hovermode='closest')
}