krz/omaha-metro-blotter
Archive of police activity and ALPR surveillance across the Omaha metro.
clone: git clone https://gitbay.org/krz/omaha-metro-blotter.git
752a4835626d3138238de6d2f896be840a2896ac
verified · cmc
author: Christian Cleberg <hello@cleberg.net> · 2024-01-24T19:34:59Z
notebooks/db_exploration.ipynb | 138 +++++++++++++++++++++++++++++++++++------ 1 file changed, 119 insertions(+), 19 deletions(-) @@ -16,14 +16,24 @@ "Let\"s explore the data a little bit to see what kind of analysis and visualizations we want to implement." ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Set up environment\n", + "\n", + "Start by installating and importing the necessary packages. " + ] + }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ - "!pip3 install ipykernel\n", - "!pip3 install --upgrade pandas plotly dash \"nbformat>=4.2.0\"" + "# Install packages, if needed\n", + "# !pip3 install ipykernel\n", + "# !pip3 install --upgrade pandas plotly dash \"nbformat>=4.2.0\"" ] }, { @@ -32,8 +42,22 @@ "metadata": {}, "outputs": [], "source": [ + "# Import packages\n", "import pandas as pd\n", - "import sqlite3" + "import numpy as np\n", + "import sqlite3\n", + "import plotly.express as px\n", + "import plotly.graph_objects as go\n", + "import plotly.io as pio" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Load Data\n", + "\n", + "To load the data, we need to connect to the SQLite3 database file and query it for the data we want." ] }, { @@ -42,7 +66,9 @@ "metadata": {}, "outputs": [], "source": [ - "connection = sqlite3.connect(\"../raw_data/ingress.db\")" + "# Connect to the database\n", + "connection = sqlite3.connect(\"../raw_data/ingress.db\")\n", + "cursor = connection.cursor()" ] }, { @@ -51,7 +77,9 @@ "metadata": {}, "outputs": [], "source": [ - "cursor = connection.cursor()" + "# If exists, delete extra header rows\n", + "# delete_headers = \"DELETE FROM incidents WHERE rb = 'RB Number'\"\n", + "# cursor.execute(delete_headers)" ] }, { @@ -60,8 +88,19 @@ "metadata": {}, "outputs": [], "source": [ - "# Test query to see if the data loaded\n", - "select_all = \"SELECT * FROM incidents;\"" + "# Grab all data\n", + "select_all = \"SELECT * FROM incidents\"\n", + "df = pd.read_sql_query(select_all, connection)\n", + "df.head()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Data Cleaning\n", + "\n", + "We will clean up the data before we use: inserting NaN, converting types, etc." ] }, { @@ -70,10 +109,31 @@ "metadata": {}, "outputs": [], "source": [ - "df = pd.read_sql_query(select_all, connection)\n", + "# Replace empty cells in [lat, lon] with NaN\n", + "df = df.replace(r'^\\s*$', np.nan, regex=True)\n", "df.head()" ] }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "# Convert date col to datetime format\n", + "df[\"date\"] = pd.to_datetime(df[\"date\"])\n", + "df" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Plotting\n", + "\n", + "Let's test a plot that will show us the top categories of incidents." + ] + }, { "cell_type": "code", "execution_count": null, @@ -81,20 +141,43 @@ "outputs": [], "source": [ "# test plotting by sorting & plotting top 5 crime categories\n", - "s = df.value_counts(subset=[\"description\"])\n", + "s = dff.value_counts(subset=[\"description\"])\n", "t = s.nlargest(5)\n", "t.head()\n", "t.plot(kind=\"bar\", title=\"Top 5 Incident Categories\")" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Data Filtering\n", + "\n", + "To reduce the workload in this rest of this notebook, I am filtering just for one description and a range of dates.\n", + "\n", + "If you are doing a lot of analysis, I recommend modifying the query at the beginning to only the pull the data you need instead of filtering after querying." + ] + }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ - "import plotly.express as px\n", - "import plotly.graph_objects as go" + "# Create a smaller dataframe based on a selected date and description\n", + "start_date = \"2023-01-01\"\n", + "end_date = \"2023-12-31\"\n", + "description = \"INJURY\"\n", + "\n", + "dff = df[(df['date'] > start_date) & (df['date'] < end_date)]\n", + "dff = dff.reset_index()\n", + "dff = dff[dff.description == description]\n", + "\n", + "dff_grouped = dff.groupby(by=\"date\").count()\n", + "dff_grouped = dff_grouped.reset_index()\n", + "\n", + "print(dff.head())\n", + "print(dff_grouped.head())" ] }, { @@ -103,7 +186,7 @@ "metadata": {}, "outputs": [], "source": [ - "s.head(10)" + "dff.size" ] }, { @@ -112,8 +195,16 @@ "metadata": {}, "outputs": [], "source": [ - "filtered_df = df[(df['date'] > '01/01/2023') & (df['date'] < '12/31/2023')]\n", - "filtered_df.head()" + "dff.info()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Mapping\n", + "\n", + "Let's create a geo map of the crime data." ] }, { @@ -123,7 +214,7 @@ "outputs": [], "source": [ "fig = px.scatter_mapbox(\n", - " filtered_df,\n", + " dff,\n", " lat=\"lat\",\n", " lon=\"lon\",\n", " color=\"description\",\n", @@ -138,7 +229,7 @@ "fig.update_layout(mapbox_style=\"open-street-map\")\n", "fig.update_layout(margin={\"r\": 0, \"t\": 0, \"l\": 0, \"b\": 0})\n", "fig.update_layout(mapbox_bounds={\"west\": -180, \"east\": -50, \"south\": 20, \"north\": 90})\n", - "# fig.show()" + "fig.show()" ] }, { @@ -147,8 +238,17 @@ "metadata": {}, "outputs": [], "source": [ - "import plotly.io as pio\n", - "pio.write_html(fig, file=\"test.html\", auto_open=False)" + "# Optionally, save the figure to an HTML file\n", + "# pio.write_html(fig, file=\"test.html\", auto_open=True)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Wrapping Up\n", + "\n", + "To finish, remember to close your database connections and save any data you need." ] }, { @@ -157,7 +257,7 @@ "metadata": {}, "outputs": [], "source": [ - "# clean up and close it out\n", + "# clean up and close out the database\n", "connection.commit()\n", "connection.close()" ]