better & updated visualizations

code/myriad/loam_paper/visualization/script.py

@@ -1,49 +1,47 @@
 import json
 import pandas as pd
 import matplotlib.pyplot as plt
-import seaborn as sns
 
 repos_info = pd.read_json("../dataset/repos_info.json")
 all_loggedin_contributors = pd.read_json("../dataset/all_loggedin_contributors.json")
 
 with open("../dataset/categories_info.json", "r") as f:
 	categories_info = json.load(f)
 category_names = [c["category"] for c in categories_info]
-category_names.sort()
 
 def avg_contributors_per_category_barplot():
     # average number of anon/loggedin contributors per category
     data = []
     for cat_info in categories_info:
         category = cat_info["category"]
         repos = repos_info.loc[repos_info["category"] == category]
-        avg_anonymous = repos.loc[:, "anonymous_contributors"].mean()
         avg_loggedin = repos.loc[:, "loggedin_contributors"].mean()
-        if avg_anonymous == avg_anonymous: data.append(
-            {
+        avg_anonymous = repos.loc[:, "anonymous_contributors"].mean()
+        if avg_anonymous is not None: 
+            data.append({
                 "category": category,
                 "logged-in": avg_loggedin,
                 "anonymous": avg_anonymous
-            }
-        )
+            })
     data = pd.DataFrame(data)
-    data.set_index("category").plot(kind="bar", stacked=False, color=["steelblue", "red"])
-    plt.yscale("log")
-    plt.xlabel("category")
-    plt.ylabel("average number of contributors")
-    plt.show()
 
-def contributions_years_scatterplot():
-    # relationship between date created and number of contributions
-    g = sns.relplot(
-        data=repos_info,
-        x="created_at", y="total_contributions",
-        hue="category", size=10
-    )
-    g.set(yscale="log")
-    plt.xticks(rotation=90)
-    plt.xlabel("year of creation/apparition on GitHub")
-    plt.ylabel("total number of contributions")
+    fig = plt.figure()
+    ax = fig.add_subplot(111)
+    ax2 = ax.twinx()
+
+    data.set_index("category").plot(kind="bar", color=["red", "orange"], stacked=False, ax=ax, width=0.5, position=0.5)
+    x = [cat["category"] for cat in categories_info]
+    y = [len(cat["repos"]) for cat in categories_info]
+    second_color = "#377eb8"
+    ax2.plot(x, y, ".", markersize=15)
+
+    ax.set_ylabel("average number of contributors")
+    ax.set_yscale("log")
+    ax.legend(title="type of contributor", loc="upper left")
+    ax2.set_ylabel("number of repos", color=second_color)
+    ax2.tick_params(axis="y", colors=second_color)
+    ax2.spines["right"].set(color=second_color, linewidth=2)
+
     plt.show()
 
 def repos_per_year_barplot():
@@ -72,16 +70,27 @@ def multi_contributors_barplot():
     idx = 0
     while all_loggedin_contributors.iloc[idx]["type"] == "Bot":
         idx+=1
-    data = [{"number of repos contributed to": i+1, "number of accounts": 0} for i in range(len(all_loggedin_contributors.iloc[idx]["contributions"]))]
+    data = [{"number of repos contributed to": i+1, "one category": 0, "multiple categories": 0} for i in range(len(all_loggedin_contributors.iloc[idx]["contributions"]))]
     for index, row in all_loggedin_contributors.iterrows():
         if (row["type"] == "Bot"): continue
-        nb_repos_contributed_to = len(row["contributions"])-1
-        data[nb_repos_contributed_to]["number of accounts"] += 1
-    data = pd.DataFrame(data)
-    g = sns.barplot(data, x="number of repos contributed to", y="number of accounts", color="orange")
-    g.bar_label(g.containers[0])
-    plt.yscale("log")
+        contributions = row["contributions"]
+        categories = []
+        for contrib in contributions:
+            for cat in categories_info:
+                if contrib["repo_name"] in cat["repos"]:
+                    categories.append(cat["category"])
+                    break
+        categories = list(set(categories))
+        nb_repos_contributed_to = len(contributions)-1
+        if len(categories) == 1:
+            data[nb_repos_contributed_to]["one category"] += 1
+        else:
+            data[nb_repos_contributed_to]["multiple categories"] += 1
+    data = pd.DataFrame(data[1:])
+    data.set_index("number of repos contributed to").plot(kind="bar", stacked=False)
     plt.ylabel("user count")
+    plt.yscale("log")
+    plt.legend(title="type of contributions")
     plt.show()
 
 def exclusive_repos_per_artwork_barplot():
@@ -91,7 +100,7 @@ def exclusive_repos_per_artwork_barplot():
     artworks.remove("loam")
     artworks.sort()
     data = []
-    for a in ["A ... past", "Apocryph", "Dear Ai", "Glommen", "... Sand Sorter", "Megatouch", "Pain", "RELAX...", "Why ... this?"]:
+    for a in ["ANSDP", "A", "DA", "G", "ISS", "M", "PC", "R", "WAIST"]:
         o = {"exclusivity": a}
         for cat in category_names:
             o[cat] = 0
@@ -103,14 +112,13 @@ def exclusive_repos_per_artwork_barplot():
         category = row["category"]
         data[idx][category] += 1
     data = pd.DataFrame(data)
-    data.set_index("exclusivity").plot(kind="bar", stacked=True)
+    data.set_index("exclusivity").plot(kind="bar", stacked=True, cmap="Set1")
     plt.xlabel("artworks")
     plt.ylabel("number of exclusive repositories")
     plt.legend(title="category")
     plt.show()
 
 avg_contributors_per_category_barplot()
-#contributions_years_scatterplot()
-repos_per_year_barplot()
+#repos_per_year_barplot()
 multi_contributors_barplot()
 exclusive_repos_per_artwork_barplot()