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