diff --git a/schema.py b/schema.py
new file mode 100644
index 0000000..82c98b1
--- /dev/null
+++ b/schema.py
@@ -0,0 +1,141 @@
+import config as app_config
+import json
+
+bucket_name = app_config.backend['s3']['name']
+
+class Owntracks(object):
+    def __init__(self, json_body):
+        """User supplies a JSON body and we verify
+        it is Owntracks app data.
+
+        Can generate responses in different formats.
+        """
+
+        # The glorious "verification" of Owntracks data
+        if "_type" not in json_body:
+            # TODO better logging
+            print("[DEBUG] Not a valid Owntracks JSON object")
+            exit(1)
+
+        self.schema_type = json_body["_type"]
+        self.schema = schemas[self.schema_type]
+        self.data = {}
+
+        required_keys = sorted(self.schema)
+        returned_keys = sorted(json_body)
+
+        # Generate new emptied dict object from default schema
+        for key in required_keys:
+            self.data[key] = ""
+        # Overwrite with the values we have
+        for key in returned_keys:
+            self.data[key] = json_body[key]
+
+        print("[DEBUG] Generating new object: {}".format(self.data))
+
+
+    def to_json(self):
+        return json.dumps(self.data)
+
+
+    def to_csv(self):
+        """Convert the JSON object to CSV"""
+        csv_data = []
+        for key in sorted(self.data):
+            csv_data.append(str(self.data[key]))
+
+        output = ",".join(csv_data)
+        return output
+
+
+    def generate_athena_query(self):
+        """Do stuff"""
+        table_tmpl = "`{name}` {type}"
+        table_columns = []
+        # table_columns = ""
+
+        for key, value in self.schema.items():
+            # table_columns += table_tmpl.format(
+            #     name=key, type=value
+            # )
+            table_columns.append(table_tmpl.format(
+                name=key, type=value
+            ))
+
+        query = create_table.format(
+            bucket_name=bucket_name,
+            table=self.schema_type,
+            table_columns=",\n".join(table_columns)
+        )
+
+        return str(query)
+
+schemas = {
+    "location": {
+        "_type": "string",
+        "acc": "int",
+        "alt": "int",
+        "batt": "int",
+        "cog": "int",
+        "conn": "string",
+        "cp": "string",
+        "datestamp": "string",
+        "device_name": "string",
+        "inregions": "string",
+        "lat": "double",
+        "lon": "double",
+        "p": "double",
+        "rad": "int",
+        "t": "string",
+        "tid": "string",
+        "timestamp": "int",
+        "topic": "string",
+        "tst": "int",
+        "vac": "int",
+        "vel": "int"
+    },
+    "beacon": {},
+    "card": {},
+    "cmd": {},
+    "configuration": {},
+    "encrypted": {},
+    "lwt": {},
+    "steps": {},
+    "transition": {
+        "acc": "double",
+        "desc": "string",
+        "event": "string",
+        "lat": "double",
+        "lon": "double",
+        "t": "string",
+        "tid": "string",
+        "tst": "int",
+        "wtst": "int"
+    },
+    "waypoint": {
+        "desc": "string",
+        "lat": "double",
+        "lon": "double",
+        "major": "int",
+        "minor": "int",
+        "rad": "int",
+        "tst": "int",
+        "uuid": "string"
+    },
+    "waypoints": {}
+}
+
+create_table = """
+CREATE EXTERNAL TABLE `{bucket_name}-{table}`(
+    {table_columns}
+)
+
+ROW FORMAT DELIMITED
+  FIELDS TERMINATED BY ','
+STORED AS INPUTFORMAT
+  'org.apache.hadoop.mapred.TextInputFormat'
+OUTPUTFORMAT
+  'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat'
+LOCATION
+  's3://{bucket_name}/{table}/'
+"""