Add analysis of time-til-first-comment for issues.

Author: rossbar

site/numpy_issues.md

@@ -119,9 +119,7 @@ p.yaxis.axis_label = TeX(r"\frac{issues}{day}")
 show(p)
 ```
 
-% TODO: add distribution of labels
-
-### First responders
+#### Time to response
 
 ```{code-cell} ipython3
 ---
@@ -135,8 +133,14 @@ newly_created_day_old = [
 ]
 
 # TODO: really need pandas here
-first_commenters = []
-for iss in newly_created_day_old:
+commented_issues = [
+    iss for iss in newly_created_day_old
+    if any(
+        e["node"]["__typename"] == "IssueComment" for e in iss["timelineItems"]["edges"]
+    )
+]
+first_commenters, time_to_first_comment = [], []
+for iss in commented_issues:
     for e in iss["timelineItems"]["edges"]:
         if e["node"]["__typename"] == "IssueComment":
             try:
@@ -145,7 +149,26 @@ for iss in newly_created_day_old:
                 # This can happen e.g. when a user deletes their GH acct
                 user = "UNKNOWN"
             first_commenters.append(user)
+            dt = np.datetime64(e["node"]["createdAt"]) - np.datetime64(iss["createdAt"])
+            time_to_first_comment.append(dt.astype("m8[m]"))
             break  # Only want the first commenter
+time_to_first_comment = np.array(time_to_first_comment)  # in minutes
+
+median_time_til_first_response = np.median(time_to_first_comment.astype(int) / 60)
+
+cutoffs = [
+    np.timedelta64(1, "h"),
+    np.timedelta64(12, "h"),
+    np.timedelta64(24, "h"),
+    np.timedelta64(3, "D"),
+    np.timedelta64(7, "D"),
+    np.timedelta64(14, "D"),
+]
+num_issues_commented_by_cutoff = np.array(
+    [
+        np.sum(time_to_first_comment < cutoff) for cutoff in cutoffs
+    ]
+)
 
 # TODO: Update IssueComment query to include:
 #  - whether the commenter is a maintainer
@@ -154,16 +177,53 @@ for iss in newly_created_day_old:
 # maintainers vs. non-maintainer, and the distribution of how long an issue
 # usually sits before it's at least commented on
 
-glue("new_issues_at_least_1_day_old", len(newly_created_day_old))
 glue(
     "num_new_issues_responded",
-    percent_val(len(first_commenters), len(newly_created_day_old))
+    percent_val(len(commented_issues), len(newly_created_day_old))
 )
+
+glue("new_issues_at_least_1_day_old", len(newly_created_day_old))
+glue("median_response_time", f"{median_time_til_first_response:1.0f}")
 ```
 
 Of the {glue:text}`new_issues_at_least_1_day_old` issues that are at least 24
 hours old, {glue:text}`num_new_issues_responded` of them have been commented
 on.
+The median time until an issue is first responded to is
+{glue:text}`median_response_time` hours.
+
+```{code-cell} ipython3
+---
+tags: [hide-input]
+---
+from bokeh.transform import dodge
+
+title = f"Percentage of issues opened since {query_date} that are commented on within..."
+
+x = [str(c) for c in cutoffs]
+y = 100 * num_issues_commented_by_cutoff / len(newly_created_day_old)
+
+p = figure(
+    x_range=x,
+    y_range=(0, 100),
+    width=670,
+    height=400,
+    title=title,
+    tooltips=[(r"%", "@top")],
+)
+p.vbar(x=x, top=y, width=0.8)
+p.xaxis.axis_label = "Time interval"
+p.yaxis.axis_label = "Percentage of issues commented on within interval"
+show(p)
+```
+
+#### First responders
+
+```{code-cell} ipython3
+---
+tags: [hide-input]
+---
+```
 
 ```{code-cell} ipython3
 ---