adding cohort retention graphs and restoring unknown traders file

app.py

@@ -1,5 +1,6 @@
 import gradio as gr
 import pandas as pd
+import seaborn as sns
 import duckdb
 import logging
 
@@ -212,7 +213,7 @@ with demo:
             )
 
             with gr.Row():
-                gr.Markdown("# Weekly metrics of  🌊 Olas traders")
+                gr.Markdown("# Weekly metrics of 🌊 Olas traders")
             with gr.Row():
                 trader_o_details_selector = gr.Dropdown(
                     label="Select a weekly trader metric",
@@ -431,21 +432,79 @@ with demo:
                         wow_retention=wow_retention
                     )
 
-            # with gr.Row():
-            #     gr.Markdown("# Cohort retention in pearl traders")
-            # with gr.Row():
-            #     cohort_retention = calculate_cohort_retention(df=retention_df)
-            #     cohort_retention_plot = plot_cohort_retention_heatmap(
-            #         retention_matrix=cohort_retention
-            #     )
-            # with gr.Row():
-            #     gr.Markdown("# Cohort retention in qs traders")
-            # with gr.Row():
-            #     cohort_retention = calculate_cohort_retention(df=retention_df)
-            #     cohort_retention_plot = plot_cohort_retention_heatmap(
-            #         retention_matrix=cohort_retention
-            #     )
-
+            with gr.Row():
+                gr.Markdown("# Cohort retention in pearl traders")
+            with gr.Row():
+                with gr.Column(scale=1):
+                    gr.Markdown("## Cohort retention of 🌊 Olas traders")
+                    cohort_retention_olas_pearl = calculate_cohort_retention(
+                        df=retention_df, market_creator="pearl", trader_type="Olas"
+                    )
+                    cohort_retention_plot1 = plot_cohort_retention_heatmap(
+                        retention_matrix=cohort_retention_olas_pearl, cmap="Purples"
+                    )
+                with gr.Column(scale=1):
+                    gr.Markdown("## Cohort retention of Non-Olas traders")
+                    # non_Olas
+                    cohort_retention_non_olas_pearl = calculate_cohort_retention(
+                        df=retention_df, market_creator="pearl", trader_type="non_Olas"
+                    )
+                    cohort_retention_plot2 = plot_cohort_retention_heatmap(
+                        retention_matrix=cohort_retention_non_olas_pearl,
+                        cmap=sns.color_palette("light:goldenrod", as_cmap=True),
+                    )
+            with gr.Row():
+                with gr.Column(scale=1):
+                    gr.Markdown("## Cohort retention of unclassified traders")
+                    cohort_retention_unclassified_pearl = calculate_cohort_retention(
+                        df=retention_df,
+                        market_creator="pearl",
+                        trader_type="unclassified",
+                    )
+                    cohort_retention_plot3 = plot_cohort_retention_heatmap(
+                        retention_matrix=cohort_retention_unclassified_pearl,
+                        cmap="Greens",
+                    )
+                with gr.Column(scale=1):
+                    print("Adding explanatory text")
+            with gr.Row():
+                gr.Markdown("# Cohort retention in quickstart traders")
+            with gr.Row():
+                with gr.Column(scale=1):
+                    gr.Markdown("## Cohort retention of 🌊 Olas traders")
+                    cohort_retention_olas_qs = calculate_cohort_retention(
+                        df=retention_df, market_creator="quickstart", trader_type="Olas"
+                    )
+                    cohort_retention_plot4 = plot_cohort_retention_heatmap(
+                        retention_matrix=cohort_retention_olas_qs,
+                        cmap="Purples",
+                    )
+                with gr.Column(scale=1):
+                    gr.Markdown("## Cohort retention of Non-Olas traders")
+                    # non_Olas
+                    cohort_retention_non_olas_qs = calculate_cohort_retention(
+                        df=retention_df,
+                        market_creator="quickstart",
+                        trader_type="non_Olas",
+                    )
+                    cohort_retention_plot5 = plot_cohort_retention_heatmap(
+                        retention_matrix=cohort_retention_non_olas_qs,
+                        cmap=sns.color_palette("light:goldenrod", as_cmap=True),
+                    )
+            with gr.Row():
+                with gr.Column(scale=1):
+                    gr.Markdown("## Cohort retention of unclassified traders")
+                    cohort_retention_unclassified_qs = calculate_cohort_retention(
+                        df=retention_df,
+                        market_creator="quickstart",
+                        trader_type="unclassified",
+                    )
+                    cohort_retention_plot6 = plot_cohort_retention_heatmap(
+                        retention_matrix=cohort_retention_unclassified_qs,
+                        cmap="Greens",
+                    )
+                with gr.Column(scale=1):
+                    print("Adding explanatory text")
         with gr.TabItem("⚙️ Active traders"):
             with gr.Row():
                 gr.Markdown("# Active traders for all markets by trader categories")

data/unknown_traders.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:de9aa0ab5acaf1623b604d88a97fa4d2369ba490e872fbeb95f8308a870e6149
-size 210024
+oid sha256:facb6d44b0ca6896cd98108283bc9527aee15ca3ca99df7a7c364ad2fb53b172
+size 320009

notebooks/closed_markets.ipynb

@@ -36,7 +36,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": 3,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -48,7 +48,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 4,
    "metadata": {},
    "outputs": [
     {
@@ -56,38 +56,38 @@
      "output_type": "stream",
      "text": [
       "<class 'pandas.core.frame.DataFrame'>\n",
-      "RangeIndex: 118880 entries, 0 to 118879\n",
+      "RangeIndex: 117525 entries, 0 to 117524\n",
       "Data columns (total 26 columns):\n",
       " #   Column                         Non-Null Count   Dtype              \n",
       "---  ------                         --------------   -----              \n",
-      " 0   collateralAmount               118880 non-null  object             \n",
-      " 1   collateralAmountUSD            118880 non-null  object             \n",
-      " 2   collateralToken                118880 non-null  object             \n",
-      " 3   creationTimestamp              118880 non-null  datetime64[ns, UTC]\n",
-      " 4   trader_address                 118880 non-null  object             \n",
-      " 5   feeAmount                      118880 non-null  object             \n",
-      " 6   id                             118880 non-null  object             \n",
-      " 7   oldOutcomeTokenMarginalPrice   118880 non-null  object             \n",
-      " 8   outcomeIndex                   118880 non-null  object             \n",
-      " 9   outcomeTokenMarginalPrice      118880 non-null  object             \n",
-      " 10  outcomeTokensTraded            118880 non-null  object             \n",
-      " 11  title                          118880 non-null  object             \n",
-      " 12  transactionHash                118880 non-null  object             \n",
-      " 13  type                           118880 non-null  object             \n",
-      " 14  market_creator                 118880 non-null  object             \n",
-      " 15  fpmm.answerFinalizedTimestamp  78679 non-null   object             \n",
-      " 16  fpmm.arbitrationOccurred       118880 non-null  bool               \n",
-      " 17  fpmm.currentAnswer             78679 non-null   object             \n",
-      " 18  fpmm.id                        118880 non-null  object             \n",
-      " 19  fpmm.isPendingArbitration      118880 non-null  bool               \n",
-      " 20  fpmm.openingTimestamp          118880 non-null  object             \n",
-      " 21  fpmm.outcomes                  118880 non-null  object             \n",
-      " 22  fpmm.title                     118880 non-null  object             \n",
-      " 23  fpmm.condition.id              118880 non-null  object             \n",
-      " 24  creation_timestamp             118880 non-null  datetime64[ns, UTC]\n",
-      " 25  creation_date                  118880 non-null  object             \n",
+      " 0   collateralAmount               117525 non-null  object             \n",
+      " 1   collateralAmountUSD            117525 non-null  object             \n",
+      " 2   collateralToken                117525 non-null  object             \n",
+      " 3   creationTimestamp              117525 non-null  datetime64[ns, UTC]\n",
+      " 4   trader_address                 117525 non-null  object             \n",
+      " 5   feeAmount                      117525 non-null  object             \n",
+      " 6   id                             117525 non-null  object             \n",
+      " 7   oldOutcomeTokenMarginalPrice   117525 non-null  object             \n",
+      " 8   outcomeIndex                   117525 non-null  object             \n",
+      " 9   outcomeTokenMarginalPrice      117525 non-null  object             \n",
+      " 10  outcomeTokensTraded            117525 non-null  object             \n",
+      " 11  title                          117525 non-null  object             \n",
+      " 12  transactionHash                117525 non-null  object             \n",
+      " 13  type                           117525 non-null  object             \n",
+      " 14  market_creator                 117525 non-null  object             \n",
+      " 15  fpmm.answerFinalizedTimestamp  77324 non-null   object             \n",
+      " 16  fpmm.arbitrationOccurred       117525 non-null  bool               \n",
+      " 17  fpmm.currentAnswer             77324 non-null   object             \n",
+      " 18  fpmm.id                        117525 non-null  object             \n",
+      " 19  fpmm.isPendingArbitration      117525 non-null  bool               \n",
+      " 20  fpmm.openingTimestamp          117525 non-null  object             \n",
+      " 21  fpmm.outcomes                  117525 non-null  object             \n",
+      " 22  fpmm.title                     117525 non-null  object             \n",
+      " 23  fpmm.condition.id              117525 non-null  object             \n",
+      " 24  creation_timestamp             117525 non-null  datetime64[ns, UTC]\n",
+      " 25  creation_date                  117525 non-null  object             \n",
       "dtypes: bool(2), datetime64[ns, UTC](2), object(22)\n",
-      "memory usage: 22.0+ MB\n"
+      "memory usage: 21.7+ MB\n"
      ]
     }
    ],
@@ -97,7 +97,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 5,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -109,7 +109,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
+   "execution_count": 6,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -127,16 +127,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
+   "execution_count": 7,
    "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "Timestamp('2025-01-13 00:00:00')"
+       "Timestamp('2024-12-28 00:00:00')"
       ]
      },
-     "execution_count": 11,
+     "execution_count": 7,
      "metadata": {},
      "output_type": "execute_result"
     }

scripts/retention_metrics.py

@@ -64,9 +64,11 @@ def calculate_wow_retention_by_type(
 
 # Cohort Retention
 def calculate_cohort_retention(
-    df: pd.DataFrame, trader_type: str, max_weeks=12
+    df: pd.DataFrame, market_creator: str, trader_type: str, max_weeks=12
 ) -> pd.DataFrame:
-    df_filtered = df.loc[df["trader_type"] == trader_type]
+    df_filtered = df.loc[
+        (df["market_creator"] == market_creator) & (df["trader_type"] == trader_type)
+    ]
     # Get first week for each trader
     first_trades = (
         df_filtered.groupby("trader_address")
@@ -76,7 +78,7 @@ def calculate_cohort_retention(
     first_trades.columns = ["trader_address", "first_trade", "cohort_week"]
 
     # Get ordered list of unique weeks - converting to datetime for proper sorting
-    all_weeks = df["month_year_week"].unique()
+    all_weeks = df_filtered["month_year_week"].unique()
     weeks_datetime = pd.to_datetime(all_weeks)
     sorted_weeks_idx = weeks_datetime.argsort()
     all_weeks = all_weeks[sorted_weeks_idx]
@@ -86,7 +88,9 @@ def calculate_cohort_retention(
 
     # Merge back to get all activities
     cohort_data = pd.merge(
-        df, first_trades[["trader_address", "cohort_week"]], on="trader_address"
+        df_filtered,
+        first_trades[["trader_address", "cohort_week"]],
+        on="trader_address",
     )
 
     # Calculate week number since first activity

tabs/retention_plots.py

@@ -22,6 +22,7 @@ def plot_wow_retention_by_type(wow_retention):
             "retention_rate": "Retention Rate (%)",
             "trader_type": "Trader Type",
         },
+        color_discrete_sequence=["purple", "goldenrod", "green"],
     )
 
     fig.update_layout(
@@ -53,13 +54,13 @@ def plot_wow_retention_by_type(wow_retention):
     )
 
 
-def plot_cohort_retention_heatmap(retention_matrix: pd.DataFrame):
+def plot_cohort_retention_heatmap(retention_matrix: pd.DataFrame, cmap: str):
 
     # Create a copy of the matrix to avoid modifying the original
     retention_matrix = retention_matrix.copy()
 
     # Convert index to datetime and format to date string
-    retention_matrix.index = pd.to_datetime(retention_matrix.index).strftime("%Y-%m-%d")
+    retention_matrix.index = pd.to_datetime(retention_matrix.index).strftime("%a-%b %d")
 
     # Create figure and axes with specified size
     plt.figure(figsize=(12, 8))
@@ -72,7 +73,7 @@ def plot_cohort_retention_heatmap(retention_matrix: pd.DataFrame):
         data=retention_matrix,
         annot=True,  # Show numbers in cells
         fmt=".1f",  # Format numbers to 1 decimal place
-        cmap="YlOrRd",  # Yellow to Orange to Red color scheme
+        cmap=cmap,  # Yellow to Orange to Red color scheme
         vmin=0,
         vmax=100,
         center=50,