diff --git a/demo/data/README.md b/demo/data/README.md
new file mode 100644
index 000000000..d2d63ec11
--- /dev/null
+++ b/demo/data/README.md
@@ -0,0 +1,2 @@
+This folder contains processed example dataset used by the demos.
+Copyright of the dataset belongs to the original copyright holder
diff --git a/wrapper/python-example/agaricus.txt.test b/demo/data/agaricus.txt.test
similarity index 100%
rename from wrapper/python-example/agaricus.txt.test
rename to demo/data/agaricus.txt.test
diff --git a/wrapper/python-example/agaricus.txt.train b/demo/data/agaricus.txt.train
similarity index 100%
rename from wrapper/python-example/agaricus.txt.train
rename to demo/data/agaricus.txt.train
diff --git a/wrapper/python-example/featmap.txt b/demo/data/featmap.txt
similarity index 100%
rename from wrapper/python-example/featmap.txt
rename to demo/data/featmap.txt
diff --git a/demo/guide-python/REAMDE.md b/demo/guide-python/REAMDE.md
new file mode 100644
index 000000000..7eaec6155
--- /dev/null
+++ b/demo/guide-python/REAMDE.md
@@ -0,0 +1,6 @@
+XGBoost Python Feature Walkthrough
+====
+* [Basic walkthrough of wrappers](guide-python/basic.py) 
+* [Cutomize loss function, and evaluation metric](guide-python/custom_objective.py)
+* [Boosting from existing prediction](guide-python/boost_from_prediction.py)
+* [Predicting using first n trees](guide-python/predict_first_ntree.py)
diff --git a/demo/guide-python/basic_walkthrough.py b/demo/guide-python/basic_walkthrough.py
new file mode 100755
index 000000000..f542954ce
--- /dev/null
+++ b/demo/guide-python/basic_walkthrough.py
@@ -0,0 +1,70 @@
+#!/usr/bin/python
+import sys
+import numpy as np
+import scipy.sparse
+# append the path to xgboost, you may need to change the following line
+# alternatively, you can add the path to PYTHONPATH environment variable
+sys.path.append('../../wrapper')
+import xgboost as xgb
+
+### simple example
+# load file from text file, also binary buffer generated by xgboost
+dtrain = xgb.DMatrix('../data/agaricus.txt.train')
+dtest = xgb.DMatrix('../data/agaricus.txt.test')
+
+# specify parameters via map, definition are same as c++ version
+param = {'max_depth':2, 'eta':1, 'silent':1, 'objective':'binary:logistic' }
+
+# specify validations set to watch performance
+watchlist  = [(dtest,'eval'), (dtrain,'train')]
+num_round = 2
+bst = xgb.train(param, dtrain, num_round, watchlist)
+
+# this is prediction
+preds = bst.predict(dtest)
+labels = dtest.get_label()
+print ('error=%f' % (  sum(1 for i in range(len(preds)) if int(preds[i]>0.5)!=labels[i]) /float(len(preds))))
+bst.save_model('0001.model')
+# dump model
+bst.dump_model('dump.raw.txt')
+# dump model with feature map
+bst.dump_model('dump.nice.txt','../data/featmap.txt')
+
+# save dmatrix into binary buffer
+dtest.save_binary('dtest.buffer')
+bst.save_model('xgb.model')
+# load model and data in 
+bst2 = xgb.Booster(model_file='xgb.model')
+dtest2 = xgb.DMatrix('dtest.buffer')
+preds2 = bst2.predict(dtest2)
+# assert they are the same
+assert np.sum(np.abs(preds2-preds)) == 0
+
+###
+# build dmatrix from scipy.sparse
+print ('start running example of build DMatrix from scipy.sparse')
+labels = []
+row = []; col = []; dat = []
+i = 0
+for l in open('../data/agaricus.txt.train'):
+    arr = l.split()
+    labels.append( int(arr[0]))
+    for it in arr[1:]:
+        k,v = it.split(':')
+        row.append(i); col.append(int(k)); dat.append(float(v))
+    i += 1
+csr = scipy.sparse.csr_matrix( (dat, (row,col)) )
+dtrain = xgb.DMatrix( csr )
+dtrain.set_label(labels)
+watchlist  = [(dtest,'eval'), (dtrain,'train')]
+bst = xgb.train( param, dtrain, num_round, watchlist )
+
+print ('start running example of build DMatrix from numpy array')
+# NOTE: npymat is numpy array, we will convert it into scipy.sparse.csr_matrix in internal implementation,then convert to DMatrix
+npymat = csr.todense()
+dtrain = xgb.DMatrix( npymat)
+dtrain.set_label(labels)
+watchlist  = [(dtest,'eval'), (dtrain,'train')]
+bst = xgb.train( param, dtrain, num_round, watchlist )
+
+
diff --git a/demo/guide-python/boost_from_prediction.py b/demo/guide-python/boost_from_prediction.py
new file mode 100755
index 000000000..0aa2e56ab
--- /dev/null
+++ b/demo/guide-python/boost_from_prediction.py
@@ -0,0 +1,26 @@
+#!/usr/bin/python
+import sys
+import numpy as np
+sys.path.append('../../wrapper')
+import xgboost as xgb
+
+dtrain = xgb.DMatrix('../data/agaricus.txt.train')
+dtest = xgb.DMatrix('../data/agaricus.txt.test')
+watchlist  = [(dtest,'eval'), (dtrain,'train')]
+###
+# advanced: start from a initial base prediction
+#
+print ('start running example to start from a initial prediction')
+# specify parameters via map, definition are same as c++ version
+param = {'max_depth':2, 'eta':1, 'silent':1, 'objective':'binary:logistic' }
+# train xgboost for 1 round
+bst = xgb.train( param, dtrain, 1, watchlist )
+# Note: we need the margin value instead of transformed prediction in set_base_margin
+# do predict with output_margin=True, will always give you margin values before logistic transformation
+ptrain = bst.predict(dtrain, output_margin=True)
+ptest  = bst.predict(dtest, output_margin=True)
+dtrain.set_base_margin(ptrain)
+dtest.set_base_margin(ptest)
+
+print ('this is result of running from initial prediction')
+bst = xgb.train( param, dtrain, 1, watchlist )
diff --git a/demo/guide-python/custom_objective.py b/demo/guide-python/custom_objective.py
new file mode 100755
index 000000000..5a7f110f4
--- /dev/null
+++ b/demo/guide-python/custom_objective.py
@@ -0,0 +1,44 @@
+#!/usr/bin/python
+import sys
+import numpy as np
+sys.path.append('../../wrapper')
+import xgboost as xgb
+###
+# advanced: cutomsized loss function
+# 
+print ('start running example to used cutomized objective function')
+
+dtrain = xgb.DMatrix('../data/agaricus.txt.train')
+dtest = xgb.DMatrix('../data/agaricus.txt.test')
+
+# note: for customized objective function, we leave objective as default
+# note: what we are getting is margin value in prediction
+# you must know what you are doing
+param = {'max_depth':2, 'eta':1, 'silent':1 }
+watchlist  = [(dtest,'eval'), (dtrain,'train')]
+num_round = 2
+
+# user define objective function, given prediction, return gradient and second order gradient
+# this is loglikelihood loss
+def logregobj(preds, dtrain):
+    labels = dtrain.get_label()
+    preds = 1.0 / (1.0 + np.exp(-preds))
+    grad = preds - labels
+    hess = preds * (1.0-preds)
+    return grad, hess
+
+# user defined evaluation function, return a pair metric_name, result
+# NOTE: when you do customized loss function, the default prediction value is margin
+# this may make buildin evalution metric not function properly
+# for example, we are doing logistic loss, the prediction is score before logistic transformation
+# the buildin evaluation error assumes input is after logistic transformation
+# Take this in mind when you use the customization, and maybe you need write customized evaluation function
+def evalerror(preds, dtrain):
+    labels = dtrain.get_label()
+    # return a pair metric_name, result
+    # since preds are margin(before logistic transformation, cutoff at 0)
+    return 'error', float(sum(labels != (preds > 0.0))) / len(labels)
+
+# training with customized objective, we can also do step by step training
+# simply look at xgboost.py's implementation of train
+bst = xgb.train(param, dtrain, num_round, watchlist, logregobj, evalerror)
diff --git a/demo/guide-python/predict_first_ntree.py b/demo/guide-python/predict_first_ntree.py
new file mode 100755
index 000000000..03f327e7f
--- /dev/null
+++ b/demo/guide-python/predict_first_ntree.py
@@ -0,0 +1,22 @@
+#!/usr/bin/python
+import sys
+import numpy as np
+sys.path.append('../../wrapper')
+import xgboost as xgb
+
+### load data in do training
+dtrain = xgb.DMatrix('../data/agaricus.txt.train')
+dtest = xgb.DMatrix('../data/agaricus.txt.test')
+param = {'max_depth':2, 'eta':1, 'silent':1, 'objective':'binary:logistic' }
+watchlist  = [(dtest,'eval'), (dtrain,'train')]
+num_round = 3
+bst = xgb.train(param, dtrain, num_round, watchlist)
+
+print ('start testing prediction from first n trees')
+### predict using first 1 tree
+label = dtest.get_label()
+ypred1 = bst.predict(dtest, ntree_limit=1)
+# by default, we predict using all the trees
+ypred2 = bst.predict(dtest)
+print ('error of ypred1=%f' % (np.sum((ypred1>0.5)!=label) /float(len(label))))
+print ('error of ypred2=%f' % (np.sum((ypred2>0.5)!=label) /float(len(label))))
diff --git a/demo/guide-python/runall.sh b/demo/guide-python/runall.sh
new file mode 100755
index 000000000..6b37c68ca
--- /dev/null
+++ b/demo/guide-python/runall.sh
@@ -0,0 +1,6 @@
+#!/bin/bash
+python basic_walkthrough.py
+python custom_objective.py
+python boost_from_prediction.py
+python boost_from_prediction.py
+rm *~ *.model *.buffer 
\ No newline at end of file
diff --git a/wrapper/README.md b/wrapper/README.md
index 3f43fa629..e736b9b6a 100644
--- a/wrapper/README.md
+++ b/wrapper/README.md
@@ -2,11 +2,10 @@ Wrapper of XGBoost
 =====
 This folder provides wrapper of xgboost to other languages
 
-
 Python
 =====
 * To make the python module, type ```make``` in the root directory of project
-* Refer to the walk through example in [python-example/demo.py](python-example/demo.py)
+* Refer also to the walk through example in [demo folder](../demo/guide-python)
 
 R 
 =====
diff --git a/wrapper/python-example/README.md b/wrapper/python-example/README.md
deleted file mode 100644
index be5350dd2..000000000
--- a/wrapper/python-example/README.md
+++ /dev/null
@@ -1,3 +0,0 @@
-example to use python xgboost, the data is generated from demo/binary_classification, in libsvm format
-
-for usage: see demo.py and comments in demo.py
diff --git a/wrapper/python-example/demo.py b/wrapper/python-example/demo.py
deleted file mode 100755
index 687b491a4..000000000
--- a/wrapper/python-example/demo.py
+++ /dev/null
@@ -1,121 +0,0 @@
-#!/usr/bin/python
-import sys
-import numpy as np
-import scipy.sparse
-# append the path to xgboost, you may need to change the following line
-# alternatively, you can add the path to PYTHONPATH environment variable
-sys.path.append('../')
-import xgboost as xgb
-
-### simple example
-# load file from text file, also binary buffer generated by xgboost
-dtrain = xgb.DMatrix('agaricus.txt.train')
-dtest = xgb.DMatrix('agaricus.txt.test')
-
-# specify parameters via map, definition are same as c++ version
-param = {'max_depth':2, 'eta':1, 'silent':1, 'objective':'binary:logistic' }
-
-# specify validations set to watch performance
-evallist  = [(dtest,'eval'), (dtrain,'train')]
-num_round = 2
-bst = xgb.train(param, dtrain, num_round, evallist)
-
-# this is prediction
-preds = bst.predict(dtest)
-labels = dtest.get_label()
-print ('error=%f' % (  sum(1 for i in range(len(preds)) if int(preds[i]>0.5)!=labels[i]) /float(len(preds))))
-bst.save_model('0001.model')
-# dump model
-bst.dump_model('dump.raw.txt')
-# dump model with feature map
-bst.dump_model('dump.nice.txt','featmap.txt')
-
-# save dmatrix into binary buffer
-dtest.save_binary('dtest.buffer')
-bst.save_model('xgb.model')
-# load model and data in 
-bst2 = xgb.Booster(model_file='xgb.model')
-dtest2 = xgb.DMatrix('dtest.buffer')
-preds2 = bst2.predict(dtest2)
-# assert they are the same
-assert np.sum(np.abs(preds2-preds)) == 0
-
-###
-# build dmatrix from scipy.sparse
-print ('start running example of build DMatrix from scipy.sparse')
-labels = []
-row = []; col = []; dat = []
-i = 0
-for l in open('agaricus.txt.train'):
-    arr = l.split()
-    labels.append( int(arr[0]))
-    for it in arr[1:]:
-        k,v = it.split(':')
-        row.append(i); col.append(int(k)); dat.append(float(v))
-    i += 1
-csr = scipy.sparse.csr_matrix( (dat, (row,col)) )
-dtrain = xgb.DMatrix( csr )
-dtrain.set_label(labels)
-evallist  = [(dtest,'eval'), (dtrain,'train')]
-bst = xgb.train( param, dtrain, num_round, evallist )
-
-print ('start running example of build DMatrix from numpy array')
-# NOTE: npymat is numpy array, we will convert it into scipy.sparse.csr_matrix in internal implementation,then convert to DMatrix
-npymat = csr.todense()
-dtrain = xgb.DMatrix( npymat)
-dtrain.set_label(labels)
-evallist  = [(dtest,'eval'), (dtrain,'train')]
-bst = xgb.train( param, dtrain, num_round, evallist )
-
-###
-# advanced: cutomsized loss function
-# 
-print ('start running example to used cutomized objective function')
-
-# note: for customized objective function, we leave objective as default
-# note: what we are getting is margin value in prediction
-# you must know what you are doing
-param = {'max_depth':2, 'eta':1, 'silent':1 }
-
-# user define objective function, given prediction, return gradient and second order gradient
-# this is loglikelihood loss
-def logregobj(preds, dtrain):
-    labels = dtrain.get_label()
-    preds = 1.0 / (1.0 + np.exp(-preds))
-    grad = preds - labels
-    hess = preds * (1.0-preds)
-    return grad, hess
-
-# user defined evaluation function, return a pair metric_name, result
-# NOTE: when you do customized loss function, the default prediction value is margin
-# this may make buildin evalution metric not function properly
-# for example, we are doing logistic loss, the prediction is score before logistic transformation
-# the buildin evaluation error assumes input is after logistic transformation
-# Take this in mind when you use the customization, and maybe you need write customized evaluation function
-def evalerror(preds, dtrain):
-    labels = dtrain.get_label()
-    # return a pair metric_name, result
-    # since preds are margin(before logistic transformation, cutoff at 0)
-    return 'error', float(sum(labels != (preds > 0.0))) / len(labels)
-
-# training with customized objective, we can also do step by step training
-# simply look at xgboost.py's implementation of train
-bst = xgb.train(param, dtrain, num_round, evallist, logregobj, evalerror)
-
-###
-# advanced: start from a initial base prediction
-#
-print ('start running example to start from a initial prediction')
-# specify parameters via map, definition are same as c++ version
-param = {'max_depth':2, 'eta':1, 'silent':1, 'objective':'binary:logistic' }
-# train xgboost for 1 round
-bst = xgb.train( param, dtrain, 1, evallist )
-# Note: we need the margin value instead of transformed prediction in set_base_margin
-# do predict with output_margin=True, will always give you margin values before logistic transformation
-ptrain = bst.predict(dtrain, output_margin=True)
-ptest  = bst.predict(dtest, output_margin=True)
-dtrain.set_base_margin(ptrain)
-dtest.set_base_margin(ptest)
-
-print ('this is result of running from initial prediction')
-bst = xgb.train( param, dtrain, 1, evallist )