How to do sentiment analysis using NLTK?

Sentiment Analysis using NLTK – One can usually find that the amount of data generated today is unstructured. This data requires processing to generate insights that are fruitful for the future. You might come across unstructured data daily: social media posts, news articles, emails, images, search history, etc.

But to make sense of it and analyze it, you would be required to use ordinary language, which falls under the field of NLP (Natural Language Processing). It is a field which is focused on forming a natural human language that is usable for computers. But how can you use it for your programs?

NLTK stands for Natural Language Toolkit and is a Python package for NLP. It is a platform for constructing Python Programs that work with human language data. NLTK comprises text processing libraries for tokenization, parsing, stemming, tagging, making datasets and semantic reasoning. With the help of this tool kit, you can process the data for sentiment analysis.

First, let’s get to know what exactly sentiment analysis is:

Sentiment Analysis  

Nowadays, every company’s main focus is to know how their latest product performs. But, for that, it is crucial to know about what is the mindset of its audience. So, how can you manage to find this information?

This is where sentiment analysis comes into use. With the help of these tool’s assistance, you can check out what your audience has to say about the product. It helps the business in understanding what is expected from them by their audience and also helps in performing better research.

Sentiment Analysis uses various techniques to figure out the message the customer wants to convey.

Steps to Perform NLTK Sentiment Analysis 

Before starting off with the NLTK process you should keep in mind installing a few tools:

Importing and Installing: 

For beginning with the task of Sentiment Analysis using NLTK, you have to install some prerequisites. The first thing that falls under that prerequisite is NLTK.

Start with installing pip which is Python’s package manager.


With this, you will be able to install the NLTK module. Still, a few more resources will be left for installation. To get hands-on with those resources, you will be required to use ():


The NLTK will display you a download manager who will show you all the installed and available resources.  Review it once and then move forward.

A faster way of downloading particular resources straightaway from the console is by passing a list to



This step allows NLTK to download and find all the resources according to the identifier. If NLTK requires you to install some additional resources that you might have missed, you will get a LookupError with all the necessary details and instructions for downloading.



A LookupError helps specify the resource that is crucial for the requested operation. It also provides instructions for downloading using the identifier.

Compilation of Data: 

With the help of NLTK, you will be provided with various functions. You can have these functions with either few or no arguments. These will help you analyze the text meaningfully before touching machine learning capabilities. The NLTK utilities are also helpful in preparing the data for advanced analysis.

To start with this step, you will be required to have some data.

For this, start by downloading the State of the Union corpus.


Keep in mind that you have built a list of individual words with the help of corpora.words() method. But you can also use str. isalpha () to include words made from letters. If you do not follow this step, your list of words may end up with words that are mere punctuation marks.

When you go through your list, you will notice that there are plenty of little words like “a,” “the,” “of,” etc. These words are called stop words and harm the accuracy of the analysis because they occur quite often. So if you are looking for an easy way to filter them out them, follow these steps:

You will be provided with a small corpus of stop words that one can load down into a list:


First, Step up the desired language as English since the corpus being mentioned here has to stop words in various languages.

You can move on further and remove stop words from the original word list.


As all the words in the stop words list are in lowercase, the original list may use str. lower () to check any discrepancies. Otherwise, you end up having a mixed case or capitalized stop words.

Using NLTK’s VADER Sentiment Analyzer 

With NLTK, you will get a pre-trained sentiment analyzer. It is known as VADER (Valence Aware Dictionary and Sentiment Reasoner).

As it is a pretrained tool, you will be getting results at a quicker scale in comparison to all the other analyzers. VADER is a well-suited tool for social media, where you can provide short sentences with some abbreviations and slang. However, it does not provide accurate results regarding longer sentences.

For using VADER, first create this instance:




You will get a dictionary of different scores. The compound score is calculated in a quite different manner. They are not just an average but range anywhere from –1 to 1. Now you can put two different corpora to test against each other. From this, you have to load up a list of strings for making the replacement.


Keep in mind that you use have used. Strings method instead of. Words method for this. By performing this method, you will get raw tweets as strings.

As different corpora have different features, you may take the help of Python’s help feature or consult NLTK’s documentation to learn about it.

NLTK sentiment analysis: Stepping stone for Your Business Success

With this blog, the major takeaway was knowing about NLTK sentiment analysis, the vital tools you should install before starting off with it, and how to use the pre-NLTK tool called VADER. With the help of this, analyze the text and gain information about all of its properties. There are plenty of different classifiers, too, that will help you gain insight into the audience’s response to the content.

With VizRefra, you can start with NLTK sentiment analysis for your business’s growth and success. With plenty of tools and dedicated staff available we will make sure to help you out in every possible way.