From: Edward Delaporte <delaport@illinois.edu>
Date: Fri, 16 Feb 2024 20:55:53 +0000 (-0600)
Subject: Cleanup unused columns
X-Git-Url: https://us-delaporte.pagekite.me/gitweb/dotfiles.git/commitdiff_plain/f2562a307496028edbcb6c337bc5fcd69ca2ffc2

Cleanup unused columns
---

diff --git a/data/issues.ipynb b/data/issues.ipynb
index 3340aaf..7608fd7 100644
--- a/data/issues.ipynb
+++ b/data/issues.ipynb
@@ -2,17 +2,9 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": null,
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "3.12.2\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "import platform\n",
     "print(platform.python_version())"
@@ -20,7 +12,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 92,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -53,7 +45,7 @@
     "\n",
     "data_sets = []\n",
     "\n",
-    "file = data_files[6]\n",
+    "file = data_files[2]\n",
     "print(file)\n",
     "\n",
     "f = open(file, encoding='utf-8')\n",
@@ -62,7 +54,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 94,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -73,17 +65,18 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 95,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
     "closed_months = pd.to_datetime(df['closedAt'], format='ISO8601')\n",
-    "df['closed_month'] = closed_months.dt.month"
+    "df['closed_month'] = closed_months.dt.month\n",
+    "df.drop(columns=['closedAt'], inplace=True)"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 97,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -92,12 +85,13 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 98,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
     "\n",
-    "grp_df = df.groupby(['mile', 'closed_month']).count().reset_index()"
+    "grp_df = df.groupby(['mile', 'closed_month']).count().reset_index()\n",
+    "grp_df.drop(columns=['milestone', 'title', 'url'], inplace=True)"
    ]
   },
   {
@@ -106,6 +100,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
+    "\n",
     "print(grp_df)"
    ]
   },
@@ -118,7 +113,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 99,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -134,7 +129,7 @@
    "outputs": [],
    "source": [
     "plt.figure(figsize=(10, 6))\n",
-    "sns.lineplot(x='closed_month', y='url', hue='mile', data=grp_df, marker='o', ci=None)\n",
+    "sns.barplot(x='closed_month', y='url', hue='mile', data=grp_df)\n",
     "plt.xlabel('Closed Month')\n",
     "plt.ylabel('ClosedAt')\n",
     "plt.title('Stacked Line Chart: Milestone Effort by Issue Close Date')\n",