diff --git a/asciidoc/courses/gds-community-detection/modules/3-kmeans/lessons/2-kmeans/lesson.adoc b/asciidoc/courses/gds-community-detection/modules/3-kmeans/lessons/2-kmeans/lesson.adoc
index f0f2c505b..f8713a463 100644
--- a/asciidoc/courses/gds-community-detection/modules/3-kmeans/lessons/2-kmeans/lesson.adoc
+++ b/asciidoc/courses/gds-community-detection/modules/3-kmeans/lessons/2-kmeans/lesson.adoc
@@ -117,6 +117,37 @@ include::questions/1-deterministic.adoc[leveloffset=+1]
 
 include::questions/2-kmeans.adoc[leveloffset=+1]
 
+
+== Troubleshooting
+
+If you receive an error stating that the graph does not exist, you may need to recreate the `survey` graph projection from the link:/courses/gds-community-detection/3-kmeans/1-scaling/[Feature normalization^] lesson.
+
+.Recreate the graph projection
+[%collapsible]
+====
+[source,cypher]
+----
+CALL gds.graph.drop('survey', false);
+
+CALL gds.graph.project(
+  'survey',
+  'Person',
+  '*',
+  {nodeProperties:['vector']}
+);
+
+CALL gds.scaleProperties.mutate(
+  'survey',
+  {
+    nodeProperties: ['vector'],
+    scaler: 'MinMax',
+    mutateProperty: 'scaledVector'
+  }
+);
+----
+====
+
+
 [.summary]
 == Summary
 
diff --git a/asciidoc/courses/gds-community-detection/modules/3-kmeans/lessons/3-optimization/lesson.adoc b/asciidoc/courses/gds-community-detection/modules/3-kmeans/lessons/3-optimization/lesson.adoc
index 0e2478cbd..2f1a6f1af 100644
--- a/asciidoc/courses/gds-community-detection/modules/3-kmeans/lessons/3-optimization/lesson.adoc
+++ b/asciidoc/courses/gds-community-detection/modules/3-kmeans/lessons/3-optimization/lesson.adoc
@@ -27,7 +27,7 @@ You will use the *stats* mode of the algorithm as you are only interested in the
 ----
 UNWIND range(5,20) AS k
 CALL gds.beta.kmeans.stats('survey',
-  {k:k, nodeProperty:'vector', computeSilhouette: True})
+  {k:k, nodeProperty:'scaledVector', computeSilhouette: true})
 YIELD averageDistanceToCentroid, averageSilhouette
 RETURN k, averageDistanceToCentroid, averageSilhouette
 ORDER BY averageSilhouette DESC LIMIT 3;
@@ -48,7 +48,7 @@ In this example, you will use the `numberOfRestarts` value 5.
 [source,cypher]
 ----
 CALL gds.beta.kmeans.write('survey',
-  {k:5, nodeProperty:'vector',
+  {k:5, nodeProperty:'scaledVector',
    numberOfRestarts: 5,
    writeProperty:'kmeans5'})
 ----
@@ -61,6 +61,37 @@ CALL gds.beta.kmeans.write('survey',
 include::questions/1-metric.adoc[leveloffset=+1]
 include::questions/2-restarts.adoc[leveloffset=+1]
 
+
+== Troubleshooting
+
+If you receive an error stating that the graph does not exist, you may need to recreate the `survey` graph projection from the link:/courses/gds-community-detection/3-kmeans/1-scaling/[Feature normalization^] lesson.
+
+.Recreate the graph projection
+[%collapsible]
+====
+[source,cypher]
+----
+CALL gds.graph.drop('survey', false);
+
+CALL gds.graph.project(
+  'survey',
+  'Person',
+  '*',
+  {nodeProperties:['vector']}
+);
+
+CALL gds.scaleProperties.mutate(
+  'survey',
+  {
+    nodeProperties: ['vector'],
+    scaler: 'MinMax',
+    mutateProperty: 'scaledVector'
+  }
+);
+----
+====
+
+
 [.summary]
 == Summary