diff --git a/.gitignore b/.gitignore
index 2619693..5797177 100644
--- a/.gitignore
+++ b/.gitignore
@@ -6,8 +6,12 @@
 /data/geojson
 /data/tiles
 /data/labdata.key
+/data/macademia_people
 map.xml
 
+*.geojson
+*.geoJSON
+
 .DS_Store
 /data/.DS_Store
 /install/.DS_Store
diff --git a/cartograph/Config.py b/cartograph/Config.py
index 9b7a21b..744d3ed 100644
--- a/cartograph/Config.py
+++ b/cartograph/Config.py
@@ -1,3 +1,4 @@
+
 from ConfigParser import SafeConfigParser
 
 EXTERNAL_FILES = 'ExternalFiles'
diff --git a/knn.py b/knn.py
new file mode 100644
index 0000000..c9d9603
--- /dev/null
+++ b/knn.py
@@ -0,0 +1,62 @@
+import luigi
+import os
+import cartograph
+import numpy as np
+from sklearn.neighbors import KDTree
+from sklearn.metrics.pairwise import cosine_similarity
+from pprint import pprint
+
+from cartograph import Config
+from cartograph import Util
+config = Config.BAD_GET_CONFIG()
+
+
+
+peopleDict = Util.read_features(config.FILE_NAME_NUMBERED_VECS, #config people
+								config.FILE_NAME_NUMBERED_NAMES,
+								config.FILE_NAME_ARTICLE_COORDINATES)
+peopleKeys = list(peopleDict.keys())
+peopleVectors = np.array([peopleDict[vID]["vector"] for vID in peopleKeys])
+peopleNames = [peopleDict[nID]["name"] for nID in peopleKeys]
+x = [float(peopleDict[fID]["x"]) for fID in peopleKeys]
+y = [float(peopleDict[fID]["y"]) for fID in peopleKeys]
+
+
+
+interestDict = Util.read_features(config.FILE_NAME_MORE_VECS,
+								config.FILE_NAME_MORE_NAMES)
+interestKeys = list(interestDict.keys())
+interestVectors = np.array([interestDict[vID]["vector"] for vID in interestKeys])
+interestNames = [interestDict[nID]["name"] for nID in interestKeys]
+
+
+kdt = KDTree(peopleVectors, leaf_size=30)
+
+
+x_lst = []
+y_lst = []
+#knn_dict = {}
+for i in range(len(interestVectors)):
+	dist, ind = kdt.query([interestVectors[i]], k=5)
+	temp_x_lst=[]
+	temp_y_lst=[]
+	temp_name_lst = []
+	weights = []
+	for j in ind[0]:
+		temp_x_lst.append(x[j])
+		temp_y_lst.append(y[j])
+		temp_name_lst.append(peopleNames[j])
+		weights.append(float(cosine_similarity([interestVectors[i]], [peopleVectors[j]]))) #cosine similarity >0
+	#print weights
+	#weights = [1,1,1,1,1]
+	x_lst.append(np.average(temp_x_lst, axis = 0, weights = weights))
+	y_lst.append(np.average(temp_y_lst, axis = 0, weights = weights))
+	#knn_dict[interestNames[i]] = temp_name_lst
+
+
+Util.write_tsv(config.FILE_NAME_MORE_COORDINATES,
+              ("index", "x", "y"), interestKeys, x_lst, y_lst)
+
+
+#pprint(knn_dict)
+
diff --git a/workflow.py b/workflow.py
index f23d4ca..e7eb191 100644
--- a/workflow.py
+++ b/workflow.py
@@ -190,7 +190,7 @@ def run(self):
             name = featureDict[featureID]["name"]
             popularityList.append(nameDict[name])
 
-        Util.write_tsv(config.FILE_NAME_NUMBERED_POPULARITY,
+        Util.write_tsv(config.get("PreprocessingFiles", "popularity_with_id"),
                        ("id", "popularity"),
                        idList, popularityList)
 
@@ -287,7 +287,7 @@ def run(self):
         X = [float(points[k]['x']) for k in keys]
         Y = [float(points[k]['y']) for k in keys]
         maxVal = max(abs(v) for v in X + Y)
-        scaling = config.MAX_COORDINATE / maxVal
+        scaling = config.get("MapConstants", "max_coordinate") / maxVal
         X = [x * scaling for x in X]
         Y = [y * scaling for y in Y]
         Util.write_tsv(config.get("PreprocessingFiles",
@@ -434,7 +434,6 @@ def run(self):
                        range(1, len(regionList) + 1),
                        regionList)
 
-
 class CreateContours(MTimeMixin, luigi.Task):
     '''
     Make contours based on density of points inside the map