From 98939fc5518ea009f4e0f33a87c2a73d642e4943 Mon Sep 17 00:00:00 2001
From: s444380 <kamgut@st.amu.edu.pl>
Date: Sun, 27 Mar 2022 11:29:22 +0200
Subject: [PATCH] Added script downloading and splitting data

---
 download_data.sh | 8 ++++++++
 1 file changed, 8 insertions(+)
 create mode 100755 download_data.sh

diff --git a/download_data.sh b/download_data.sh
new file mode 100755
index 0000000..c636d37
--- /dev/null
+++ b/download_data.sh
@@ -0,0 +1,8 @@
+#!/bin/bash
+
+kaggle datasets download -d AnalyzeBoston/crimes-in-boston
+unzip crimes-in-boston.zip
+shuf crime.csv > crime.shuf
+head -n 30000 crime.shuf > crime.test
+head -n 60000 crime.shuf | tail -n 30000 > crime.dev
+tail -n +60001 crime.shuf > crime.train