diff --git a/api/.gitkeep b/api/.gitkeep
new file mode 100644
index 0000000..e69de29
diff --git a/docs/spec.md b/docs/spec.md
index b70db44..3f9f309 100644
--- a/docs/spec.md
+++ b/docs/spec.md
@@ -13,7 +13,7 @@ Aspirationally, it should be able to track multiple trains at any one time, with
 ## Data sources
 
 The Project will source data from multiple Network Rail Open Data Feeds:
-* [Train Movements](https://wiki.openraildata.com/index.php?title=Train_Movements) or [TD](https://wiki.openraildata.com/index.php?title=TD) feed (which one is TBC) will be used to retrieve the real-time location of trains on the Network Rail network.
+* [Train Movements](https://wiki.openraildata.com/index.php?title=Train_Movements) will be used to retrieve the real-time location of trains on the Network Rail network.
 * [SCHEDULE](https://wiki.openraildata.com/index.php?title=SCHEDULE) will be used to retrieve train scheduling information.
 
 Additionally, the project will make use of the [OpenStreeMap Overpass API](https://wiki.openstreetmap.org/wiki/Overpass_API) to source data about track layout. It may also be used to find station locations.
@@ -25,3 +25,31 @@ In addition to a Redis database, there are three main components to the project:
 * A Python API server that powers the frontend application
 * The frontend JavaScript application
 
+## Redis Database
+
+The Redis database will store:
+* Real-time train location by train ID
+* Schedule information
+* TIPLOC information (see below)
+
+## Data sources
+
+### SCHEDULE
+
+Authoritatitve scheduling information can be retrieved from Network Rail by downloading a single file. This will be parsed and loaded into the Redis database.
+
+The downloads contain information about schedules and TIPLOCs, where a TIPLOC is an identifier for a physical location on the rail network. Every schedule contains timings for when a given service will pass through a given set of TIPLOCs.
+
+Each schedule is identified with a so-called "headcode".
+
+### Train Movements
+
+The Train Movements (herein referred to as TM) feed provides real-time updates as to where a train is. It emits messages of different types, documented below:
+
+![](https://wiki.openraildata.com/images/f/fb/Trust_fsm.png)
+
+`Activation` messages link a train ID to a headcode. No other messages types do this, so if the activation message is missed we will be unable to adequately understand any other messages relating to that train.
+
+### TIPLOC Information
+
+TIPLOCs can be mapped to real-world Northings and Eastings and hence then used as 
\ No newline at end of file
diff --git a/feedprocessor/.gitkeep b/feedprocessor/.gitkeep
new file mode 100644
index 0000000..e69de29
diff --git a/web/.gitkeep b/web/.gitkeep
new file mode 100644
index 0000000..e69de29