Merge pull request #63 from boostcampaitech7/feat-034/add-dataset

feat: add last deposit variable
boostcampaitech7 · Oct 23, 2024 · 4e084e2 · 4e084e2
2 parents ef262e5 + c892742
commit 4e084e2
Showing 1 changed file with 44 additions and 6 deletions.
diff --git a/merge_data.ipynb b/merge_data.ipynb
@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 32,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -19,9 +19,16 @@
     "from sklearn.preprocessing import StandardScaler\n",
     "from utils.clustering import ClusteringMethods\n",
     "\n",
-    "from tqdm import tqdm\n"
+    "from tqdm import tqdm"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
   {
    "cell_type": "code",
    "execution_count": 2,
@@ -1314,6 +1321,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
+    "from haversine import haversine, Unit\n",
     "def calculate_subway_distances(train_data, subway_data):\n",
     "    \"\"\"\n",
     "    모든 아파트에 대해 연산하면 시간이 오래 걸리므로 위도 경도에 대해서 계산을 수행한 후 Merge한다.\n",
@@ -1458,6 +1466,27 @@
     "test_data = pd.concat([test_data1, test_data2]).sort_index()"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 직전 아파트 거래가"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 지난 아파트 거래가 변수\n",
+    "train_data = train_data.sort_values(['contract_year_month', 'contract_day'])\n",
+    "train_data['last_deposit_by_area'] = train_data.groupby(['latitude', 'longitude'])['deposit_by_area'].shift(1)\n",
+    "\n",
+    "last_apt_deposit = train_data.groupby(['latitude', 'longitude']).last().reset_index()[['latitude', 'longitude', 'deposit_by_area']].rename(columns = {'deposit_by_area' : 'last_deposit_by_area'})\n",
+    "test_data = test_data.merge(last_apt_deposit, how= 'left')"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 38,
@@ -1467,6 +1496,15 @@
     "final_df = pd.concat([train_data, test_data],axis=0)"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "final_df.drop(columns = 'index', inplace = True)"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 40,
@@ -1970,7 +2008,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "final_df.to_csv(\"final_df.csv\", index=False)"
+    "final_df.to_csv(\"real_final_df.csv\", index=False)"
    ]
   },
   {
@@ -1979,7 +2017,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "final_df.to_csv(\"final_df.csv.gz\", index=False, compression='gzip')"
+    "final_df.to_csv(\"real_final_df.csv.gz\", index=False, compression='gzip')"
    ]
   },
   {
@@ -1992,7 +2030,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "base",
+   "display_name": ".conda",
    "language": "python",
    "name": "python3"
   },
@@ -2006,7 +2044,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.13"
+   "version": "3.11.9"
   }
  },
  "nbformat": 4,