From fb6b0c21b1c6d3f37fde1def993ff91d42162e3a Mon Sep 17 00:00:00 2001
From: luke-strange <92686634+luke-strange@users.noreply.github.com>
Date: Mon, 24 Jun 2024 14:46:06 +0100
Subject: [PATCH] Add dvc tracking for youth employment

---
 output/.gitignore                       |  1 +
 pipelines/people/dvc.lock               | 20 ++++++++++++++++++++
 pipelines/people/dvc.yaml               | 12 +++++++++++-
 pipelines/people/youth_employment.ipynb | 11 ++++++++---
 4 files changed, 40 insertions(+), 4 deletions(-)
 create mode 100644 output/.gitignore

diff --git a/output/.gitignore b/output/.gitignore
new file mode 100644
index 0000000..fa65608
--- /dev/null
+++ b/output/.gitignore
@@ -0,0 +1 @@
+*.ipynb
diff --git a/pipelines/people/dvc.lock b/pipelines/people/dvc.lock
index 77de1e4..2029227 100644
--- a/pipelines/people/dvc.lock
+++ b/pipelines/people/dvc.lock
@@ -119,3 +119,23 @@ stages:
       hash: md5
       md5: 121be03fe40ba8f11044310423afd4bb
       size: 23551
+  youth-employment:
+    cmd: papermill --no-progress-bar --no-report-mode youth_employment.ipynb ../../output/youth_employment.ipynb
+    deps:
+    - path: ../../working/cs/youth-unemployment-adjusted.csv
+      hash: md5
+      md5: 666b3a41e5b413f8ab738ccdd1cb3d2a
+      size: 26728
+    - path: youth_employment.ipynb
+      hash: md5
+      md5: c7079b96e21ce303072f7b8760054459
+      size: 10505
+    outs:
+    - path: ../../src/themes/people-skills-future/_data/youth_employment_RGN.csv
+      hash: md5
+      md5: cd7f7fca9117146f7178e4e076e459f9
+      size: 2666
+    - path: ../../src/themes/people-skills-future/_data/youth_employment_RGN_bar_chart.csv
+      hash: md5
+      md5: 9f432360b6e848c5969954a0880248b6
+      size: 337
diff --git a/pipelines/people/dvc.yaml b/pipelines/people/dvc.yaml
index dacde6f..1137730 100644
--- a/pipelines/people/dvc.yaml
+++ b/pipelines/people/dvc.yaml
@@ -47,4 +47,14 @@ stages:
       - ${TOP}/src/themes/people-skills-future/_data/neet.csv:
           cache: false
       - ${TOP}/src/themes/people-skills-future/_data/risk_of_neet_by_la.csv:
-          cache: false
\ No newline at end of file
+          cache: false
+  youth-employment: 
+    cmd: papermill --no-progress-bar --no-report-mode youth_employment.ipynb ${TOP}/output/youth_employment.ipynb 
+    deps: 
+      - youth_employment.ipynb
+      - ${TOP}/working/cs/youth-unemployment-adjusted.csv
+    outs: 
+      - ${TOP}/src/themes/people-skills-future/_data/youth_employment_RGN.csv:
+          cache: false
+      - ${TOP}/src/themes/people-skills-future/_data/youth_employment_RGN_bar_chart.csv:
+          cache: false
diff --git a/pipelines/people/youth_employment.ipynb b/pipelines/people/youth_employment.ipynb
index e30ce22..6ffbd4b 100644
--- a/pipelines/people/youth_employment.ipynb
+++ b/pipelines/people/youth_employment.ipynb
@@ -1,5 +1,12 @@
 {
  "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "If you want to run the notebook locally, remove the hashtag before os.chdir('../..') below."
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 1,
@@ -7,7 +14,7 @@
    "outputs": [],
    "source": [
     "import os\n",
-    "os.chdir('../..')\n",
+    "os.chdir('../..') \n",
     "from pipelines.util import *"
    ]
   },
@@ -18,8 +25,6 @@
    "outputs": [],
    "source": [
     "data = pd.read_csv('working/cs/youth-unemployment-adjusted.csv')\n",
-    "# data['unix'] = pd.to_datetime(data['date'], format=f'%Y-%m-%d').astype(int).div(10**9).astype(int)\n",
-    "# data['decimal_date'] = data['unix'].div((86400*365.25)).add(1970).round(2)\n",
     "data = add_decimal_date_to_dataframe(data, datename='date')"
    ]
   },