From 19ceaca42b7fc91c1b9bd275bec52cad71d3da05 Mon Sep 17 00:00:00 2001
From: lukaszett <mail@lz-a.de>
Date: Mon, 25 Jul 2022 22:29:53 +0200
Subject: [PATCH] add base sourcecode

---
 src/create_dataset.py | 11 +++++++++++
 src/create_id_list.py |  8 ++++++++
 2 files changed, 19 insertions(+)
 create mode 100644 src/create_dataset.py
 create mode 100644 src/create_id_list.py

diff --git a/src/create_dataset.py b/src/create_dataset.py
new file mode 100644
index 0000000..2c90e8b
--- /dev/null
+++ b/src/create_dataset.py
@@ -0,0 +1,11 @@
+import pandas as pd
+
+tweets_path = "../raw_data/tweets.csv"
+
+tweets = pd.read_csv(tweets_path)
+
+# TODO extract mentions from tweets
+
+# TODO generate cooccurence matrix
+
+# TODO save cooccurence matrix
diff --git a/src/create_id_list.py b/src/create_id_list.py
new file mode 100644
index 0000000..5e9bcd7
--- /dev/null
+++ b/src/create_id_list.py
@@ -0,0 +1,8 @@
+import pandas as pd
+
+tweets_path = "../raw_data/tweets.csv"
+
+tweets = pd.read_csv(tweets_path)
+
+for id in tweets.id:
+    print(id)