diff --git a/changelog.md b/changelog.md
--- a/changelog.md
+++ b/changelog.md
@@ -1,3 +1,11 @@
+0.2.3
+
+* Coal dataset
+
+* New internal API
+
+* Ord instance for IrisClass
+
 0.2.2
 
 * Enum, bounded instances for IrisClass
diff --git a/datasets.cabal b/datasets.cabal
--- a/datasets.cabal
+++ b/datasets.cabal
@@ -1,5 +1,5 @@
 Name:                datasets
-Version:             0.2.2
+Version:             0.2.3
 Synopsis:            Classical data sets for statistics and machine learning
 Description:         Classical machine learning and statistics datasets from
                      the UCI Machine Learning Repository and other sources.
@@ -59,6 +59,7 @@
                  , Numeric.Datasets.Adult
                  , Numeric.Datasets.BreastCancerWisconsin
                  , Numeric.Datasets.Car
+                 , Numeric.Datasets.Coal
                  , Numeric.Datasets.CO2
                  , Numeric.Datasets.Gapminder
                  , Numeric.Datasets.Iris
diff --git a/src/Numeric/Datasets.hs b/src/Numeric/Datasets.hs
--- a/src/Numeric/Datasets.hs
+++ b/src/Numeric/Datasets.hs
@@ -16,7 +16,7 @@
 
 -}
 
-{-# LANGUAGE OverloadedStrings #-}
+{-# LANGUAGE OverloadedStrings, GADTs #-}
 
 module Numeric.Datasets where
 
@@ -27,7 +27,7 @@
 import Data.Monoid
 import qualified Data.ByteString.Lazy as BL
 import qualified Data.Vector as V
-import qualified Data.Aeson as JSON
+import Data.Aeson as JSON
 import Control.Applicative
 import Data.Time
 import Data.Char (ord)
@@ -36,6 +36,7 @@
 
 import Data.Char (toUpper)
 import Text.Read (readMaybe)
+import Data.Maybe (fromMaybe)
 import Data.ByteString.Char8 (unpack)
 import qualified Data.ByteString.Lazy.Char8 as BL8
 import Data.ByteString.Lazy.Search (replace)
@@ -45,41 +46,72 @@
 -- |Load a dataset, using the system temporary directory as a cache
 getDataset :: Dataset a -> IO [a]
 getDataset ds = do
-  dir <- getTemporaryDirectory
-  ds $ dir </> "haskds"
-
--- | A dataset is defined as a function from the caching directory to the IO action that loads the data
-type Dataset a = FilePath -- ^ Directory for caching downloaded datasets
-                 -> IO [a]
+  dir <- case temporaryDirectory ds of
+           Nothing -> getTemporaryDirectory
+           Just tdir -> return tdir
+  bs <- fmap (fromMaybe id $ preProcess ds) $ getFileFromSource dir $ source ds
+  return $ readDataset (readAs ds) bs
 
--- * Defining datasets
+-- |Read a ByteString into a Haskell value
+readDataset :: ReadAs a -> BL.ByteString -> [a]
+readDataset JSON bs =
+  case JSON.decode bs of
+    Just theData ->  theData
+    Nothing -> error "failed to parse json"
+readDataset (CSVRecord hhdr opts) bs =
+  case decodeWith opts hhdr bs of
+    Right theData -> V.toList theData
+    Left err -> error err
+readDataset (CSVNamedRecord opts) bs =
+  case decodeByNameWith opts bs of
+    Right (_,theData) -> V.toList theData
+    Left err -> error err
 
 data Source = URL String
 
+-- | A dataset is a record telling us how to load the data
+
+data Dataset a = Dataset
+  { source :: Source
+  , temporaryDirectory :: Maybe FilePath
+  , preProcess :: Maybe (BL.ByteString -> BL.ByteString)
+  , readAs :: ReadAs a
+  }
+
+-- | ReadAs is a datatype to describe data formats that hold data sets
+
+data ReadAs a where
+  JSON :: FromJSON a => ReadAs a
+  CSVRecord :: FromRecord a => HasHeader -> DecodeOptions -> ReadAs a
+  CSVNamedRecord :: FromNamedRecord a => DecodeOptions -> ReadAs a
+
+csvRecord :: FromRecord a => ReadAs a
+csvRecord = CSVRecord NoHeader defaultDecodeOptions
+
+-- * Defining datasets
+
 -- |Define a dataset from a pre-processing function and a source for a CSV file
 csvDatasetPreprocess :: FromRecord a => (BL.ByteString -> BL.ByteString) -> Source -> Dataset a
-csvDatasetPreprocess preF src cacheDir = do
-  parseCSV preF <$> getFileFromSource cacheDir src
+csvDatasetPreprocess preF src = (csvDataset src) { preProcess = Just preF }
+--  parseCSV preF <$> getFileFromSource cacheDir src
 
 -- |Define a dataset from a source for a CSV file
 csvDataset :: FromRecord a =>  Source -> Dataset a
-csvDataset  = csvDatasetPreprocess id
+csvDataset src = Dataset src Nothing Nothing $ CSVRecord NoHeader defaultDecodeOptions
 
 -- |Define a dataset from a source for a CSV file with a known header
 csvHdrDataset :: FromNamedRecord a => Source -> Dataset a
-csvHdrDataset src cacheDir = do
-  parseCSVHdr <$> getFileFromSource cacheDir src
+csvHdrDataset src = Dataset src Nothing Nothing $ CSVNamedRecord defaultDecodeOptions
 
 -- |Define a dataset from a source for a CSV file with a known header and separator
 csvHdrDatasetSep :: FromNamedRecord a => Char -> Source -> Dataset a
-csvHdrDatasetSep sepc src cacheDir = do
-  parseCSVHdrSep sepc <$> getFileFromSource cacheDir src
+csvHdrDatasetSep sepc src
+   = Dataset src Nothing Nothing
+       $ CSVNamedRecord defaultDecodeOptions { decDelimiter = fromIntegral (ord sepc)}
 
 -- |Define a dataset from a source for a JSON file -- data file must be accessible with HTTP, not HTTPS
-jsonDataset :: JSON.FromJSON a => Source -> Dataset a
-jsonDataset src cacheDir = do
-  bs <- getFileFromSource cacheDir src
-  return $ parseJSON bs
+jsonDataset :: FromJSON a => Source -> Dataset a
+jsonDataset src = Dataset src Nothing Nothing JSON
 
 -- | Get a ByteString from the specified Source
 getFileFromSource :: FilePath -> Source -> IO (BL.ByteString)
@@ -95,36 +127,6 @@
        let bs = rsp ^. Wreq.responseBody
        BL.writeFile fnm bs
        return bs
-
--- | Parse CSV file
-parseCSV :: FromRecord a => (BL.ByteString -> BL.ByteString) -> BL.ByteString -> [a]
-parseCSV preF contents =
-        case decode NoHeader (preF contents) of
-          Right theData -> V.toList theData
-          Left err -> error err
-
--- | Parse CSV file with known header
-parseCSVHdr :: FromNamedRecord a => BL.ByteString -> [a]
-parseCSVHdr contents =
-        case decodeByName contents of
-          Right (_,theData) -> V.toList theData
-          Left err -> error err
-
--- | Parse CSV file with known header
-parseCSVHdrSep :: FromNamedRecord a => Char -> BL.ByteString -> [a]
-parseCSVHdrSep sepc contents =
-        let opts = defaultDecodeOptions { decDelimiter = fromIntegral (ord sepc)} in
-        case decodeByNameWith opts contents of
-          Right (_,theData) -> V.toList theData
-          Left err -> error err
-
--- | Parse JSON file
-parseJSON :: JSON.FromJSON a => BL.ByteString -> [a]
-parseJSON bs = case JSON.decode bs of
-  Just theData ->  theData
-  Nothing -> error "failed to parse json"
-
-
 
 -- * Helper functions for parsing
 
diff --git a/src/Numeric/Datasets/Coal.hs b/src/Numeric/Datasets/Coal.hs
new file mode 100644
--- /dev/null
+++ b/src/Numeric/Datasets/Coal.hs
@@ -0,0 +1,31 @@
+{-# LANGUAGE DeriveGeneric, OverloadedStrings                   #-}
+{-# LANGUAGE GADTs, QuasiQuotes, ViewPatterns, FlexibleContexts #-}
+
+{-|
+
+Coal data set
+
+Dates of mining disasters, from the `coal` dataset in the R package `boot`.
+
+For further information, see <http://vincentarelbundock.github.io/Rdatasets/doc/boot/coal.html>
+
+-}
+
+module Numeric.Datasets.Coal ( Coal, coal, date ) where
+
+import Numeric.Datasets
+
+import Data.Csv
+import GHC.Generics
+import Control.Applicative
+
+data Coal = Coal
+  { date :: Double
+  } deriving (Show, Read, Generic)
+
+instance FromRecord Coal where
+  parseRecord v = Coal <$> v .! 1
+
+coal :: Dataset Coal
+coal = let src = URL "http://vincentarelbundock.github.io/Rdatasets/csv/boot/coal.csv"
+       in Dataset src Nothing Nothing $ CSVRecord HasHeader defaultDecodeOptions
diff --git a/src/Numeric/Datasets/Iris.hs b/src/Numeric/Datasets/Iris.hs
--- a/src/Numeric/Datasets/Iris.hs
+++ b/src/Numeric/Datasets/Iris.hs
@@ -19,7 +19,7 @@
 
 
 data IrisClass = Setosa | Versicolor | Virginica
-  deriving (Show, Read, Eq, Generic, Enum, Bounded)
+  deriving (Show, Read, Eq, Ord, Generic, Enum, Bounded)
 
 instance FromField IrisClass where
   parseField "Iris-setosa" = return Setosa
@@ -38,4 +38,4 @@
 instance FromRecord Iris
 
 iris :: [Iris]
-iris = parseCSV id $ fromStrict $(embedFile "datafiles/iris.data")
+iris = readDataset csvRecord (fromStrict $(embedFile "datafiles/iris.data"))
diff --git a/src/Numeric/Datasets/Michelson.hs b/src/Numeric/Datasets/Michelson.hs
--- a/src/Numeric/Datasets/Michelson.hs
+++ b/src/Numeric/Datasets/Michelson.hs
@@ -18,4 +18,4 @@
 
 
 michelson :: [[Double]]
-michelson = parseJSON $ fromStrict $(embedFile "datafiles/michelson.json")
+michelson = readDataset JSON (fromStrict $(embedFile "datafiles/michelson.json"))
diff --git a/src/Numeric/Datasets/Nightingale.hs b/src/Numeric/Datasets/Nightingale.hs
--- a/src/Numeric/Datasets/Nightingale.hs
+++ b/src/Numeric/Datasets/Nightingale.hs
@@ -31,4 +31,4 @@
 instance FromJSON Nightingale
 
 nightingale :: [Nightingale]
-nightingale = parseJSON $ fromStrict $(embedFile "datafiles/nightingale.json")
+nightingale = readDataset JSON $ fromStrict $(embedFile "datafiles/nightingale.json")
diff --git a/src/Numeric/Datasets/OldFaithful.hs b/src/Numeric/Datasets/OldFaithful.hs
--- a/src/Numeric/Datasets/OldFaithful.hs
+++ b/src/Numeric/Datasets/OldFaithful.hs
@@ -18,7 +18,8 @@
   { recordingDay :: Int -- ^ Recording day. 1-8: 1978, 16-23: 1979
   , waiting :: Double -- ^ waiting time until next eruption
   , duration :: Double -- ^ duration of eruption in minutes
-  }
+  } deriving Show
+
 -- The first comp
 oldFaithful :: [OldFaithful]
 oldFaithful =