diff --git a/start_docker.sh b/start_docker.sh
index 4b3a6c4..fc5b455 100755
--- a/start_docker.sh
+++ b/start_docker.sh
@@ -17,6 +17,8 @@ fi
 # stored in a separate variable for convenience.
 declare repo_path="$(cd .. && git rev-parse --show-toplevel && cd $this_dir)"
 declare repo_path_out=$repo_path
+declare spark_data_path="$HOME/sparkdata"
+declare spark_data_path_out=$spark_data_path
 # `working_dir` is the starting dir for all docker shells launched via
 # `docker exec -it <container> bash`.
 declare working_dir=$repo_path_out
@@ -30,6 +32,7 @@ docker run \
   --security-opt seccomp=unconfined \
   --name $cont_name \
   -v $repo_path:$repo_path_out \
+  -v $spark_data_path:$spark_data_path_out \
   -u $(id -u $USER) \
   -w $working_dir \
   --shm-size=4024m \