This is a sentiment analysis project in python using twitter and it's API

For Python 3.4 and above

Using the GitHub repository

Install by running these commands in command prompt:

git clone https://github.com/tweepy/tweepy.git
cd tweepy   
python setup.py install

Using pip

pip install tweepy

python setup.py 

Using pip

pip install tweepy

Standard example to pull twitter feeds based on

Streaming tweets example

Test code : Extracting only the english language tweets using twitter_streaming.py

Output of the tweet stream is in the JSON format from which the "text:" specifies the tweet which is extracted from the JSON text.

Output:

Output is stored to a output text file using the command line

python twitter_streaming.py >> output.txt 

The extraction process can be allowed to run until rate limitation by twitter occurs or alternatively the process can be interrupted using

Ctrl+c

For Python 3.4 and above

NLTK - The fastest way to install is using pip NLTK dependancies:

Numpy - Requires a GNU compiler like MinGW or Visual Studio 2013 or higher installed

• Install MinGW with C++ Compiler option checked
• Add C:\MinGW\bin to your PATH

Download and run mingw-get-inst (the download link above).

Select "Use pre-packaged repository catalogues".

Review and accept the License agreement.

Please note that MinGW should be installed to a directory path that doesn't contain any spaces.

This method has been tested with a directory path of C:\MinGW.

Select C++ Compiler and MSYS Basic System as optional components.

Wait until every package has been downloaded and installation is finished.

Go to your MSYS folder (found at \msys\1.0, C:\MinGW\msys\1.0 in this example), open etc\fstab with a text editor (for example Notepad) and add the following line at the end of the file:

C:\MinGW\   /usr/local

On Windows Vista and newer, you'll need additional steps to make MSYS fully work if User Account Control is enabled (it is by default).

Go to your MSYS folder (C:\MinGW\msys\1.0 here), open msys.bat with a text editor (right-click -> Open With -> Notepad or equivalents) and add the following line after
@echo off:

cd "C:\MinGW\msys\1.0"

And if your MSYS installation drive is not the disk Windows is installed on, add one more line:

C:(MSYS drive)

After saving the file, right click on msys.bat and choose "Run as Administrator". You will need to do this every time you run MSYS. After that, programs requiring admin rights (such as install and patch) will work. Testing MinGW/MSYS installation

Open the MinGW shell (MSYS) by running msys.bat. Run the following commands:

make -v
gcc -v

They should output something. Check if something goes wrong. Compilation and installation of the required packages

To compile and install these packages and avoid the error 'wget: command not found', first you need to install the following commands: (To compile on Windows 7, the service "application experience" must be activated and running)

IN MSYS:

mingw-get install msys-wget
mingw-get install msys-unzip
mingw-get install msys-patch

If you get following error: configure: error: cannot run C compiled programs. , one of the reasons could be your anti-virus, that silently deletes a.exe binary files. Try to temporarily disable anti-virus software and compile again

• In PYTHONPATH\Lib\distutils, create a file distutils.cfg and add these lines:

    [build]
    compiler=mingw32
  

Then install using either

pip install numpy
pip install nltk

or alternatively using the Git Repositories

git clone https://github.com/numpy/numpy.git
cd tweepy
python setup.py install
git clone https://github.com/nltk/nltk.git
cd tweepy
python setup.py install

After NLTK is installed open the Python interpreter and type

>>import nltk
>>nltk.download

In the NLTK Downloader dialog box select All packages and click the Download button

Matplotlib Dependencies:

• numpy 1.6 (or later) Already installed for NLTK

• libpng 1.2 (or later)

To be installed via MinGW If MSYS is not setup the instructions are here.

If not installed Open the MinGW setup and select and install the MSYS basic system package

libpng requires zlib to install zlib do the following,

In MSYS:

    wget http://zlib.net/zlib-1.2.8.tar.gz
    tar xvfz zlib-1.2.8.tar.gz
    cd zlib-1.2.8
    make -f win32/Makefile.gcc BINARY_PATH=/usr/local/bin INCLUDE_PATH=/usr/local/include LIBRARY_PATH=/usr/local/lib install
    cd ..

then proceed with

    wget http://sourceforge.net/projects/libpng/files/libpng15/older-releases/1.5.16/libpng-1.5.16.tar.xz/download
    tar xvfJ libpng-1.5.16.tar.xz
    cd libpng-1.5.16
    mv INSTALL INSTALL.txt
    ./configure
    make install
    cd ..

• freetype 2.3 or later

To be installed via MinGW

In MSYS:

    wget http://sourceforge.net/projects/libpng/files/libpng15/older-releases/1.5.16/libpng-1.5.16.tar.xz/download
    tar xvfJ libpng-1.5.16.tar.xz
    cd libpng-1.5.16
    mv INSTALL INSTALL.txt
    ./configure
    make install
    cd ..

• dateutil 1.1 or later

If using pip, easy_install or installing from source, the installer will attempt to download and install python_dateutil from PyPI. Note that python_dateutil also depends on six. pip and other package managers should handle installing that secondary dependency automatically.

• pyparsing

If using pip, easy_install or installing from source, the installer will attempt to download and install pyparsing from PyPI.

• six 1.4 or later

Also a dependency of dateutil.

• pytz

Test Code: To evaluate our twitter tweets positive and negative sentiments we implement tweets_score.py. It utilises a Patricia Trie constructed using a python dictionary data structure with wordbanks negative-words.txt and positive-words.txt as data. Using these tries we optimise the wordbank search for positive and negative word matches. Test Tweet data used in this example is here.

Output

positive = 78 negative = 20 unknown = 120

Shown by a plot using pythons Matplotlib