1 of 13

1

Organization:

Narcotics Control Bureau (NCB)

Problem Statement Title:

Crawling of the Darknet markets to identify drug traffickers on dark net based in India and the drugs they offer for sale.

Category:

Software

PS Number:

PK897

Domain Bucket:

Miscellaneous

The Nasty Crawlers

2 of 13

2

India lies sandwiched between two of the world’s “three key production areas” for opium, leading to an illicit drugs crisis. But cross-border trafficking holds only part of the explanation for the illicit drugs trade that runs through India. India’s pharmaceutical industry is “the largest provider of generic drugs globally”, as per the department of pharmaceuticals, Ministry of Chemicals and Fertilizers. But there have been reports of clandestine manufacturing and siphoning of synthetic drugs, as well as illicit trade in chemical precursors.

Currently there is no digital way of reporting drug trafficking. The only way is to reach out Narcotics squad or nearby police station. However, this is not feasible since majority of public will not be ready to disclose their profile or will not be ready to spend their time physically. The intended solution should enable public to identify drug traffickers on dark net based in India and the drugs they offer for sale. The solution should enable the public to upload any relevant information about the finding and it should be easy for the user to use with different language supports.

Problem Statement

3 of 13

3

Drugs have become one of the major concerns in youth now a days. The consumption of drugs have increased in India four-fold time from 2017 to 2020. As most of these drugs are sold on the dark web, it is very difficult to identify and locate these drug peddlers. So an autonomous dark web crawler can prove to be very helpful in identifying the sources of these drug operators.

Need of Solution

4 of 13

4

The main goal of this project is to gather open data from the deep web (also known as the dark web) and, using data mining methods, compile as much information as possible into an interactive tree graph. The interactive tree graph module will be able to demonstrate the relationships between the intelligence data that has been acquired.

Need of Solution

5 of 13

5

A Web crawler that:

Crawls through the dark web
Work on a directory of links to crawl
Take them as seed to enter into the dark web
Scrape the collected data for links and continue.
Scrapes through the collected data to return a list of Indian drug marketplaces.
Analyse the Content and DOM structure to check for drug marketplace.

Proposed Solution

6 of 13

6

Methodology - Architecture Diagram

7 of 13

7

Methodology - Activity Diagram

8 of 13

Crawls only Drug Related pages�
Does both structure as well as content analysis of pages�
Uses Name Entity Recognition (N.E.R) to analyse and identify the relevant drug marketplaces �
Support for multiple languages

Features/Use Cases of Proposed Solution

9 of 13

9

The application will be build majorly using Python and its vast libraries.

LSTM (DOM analysis and Content analysis) - NumPy, Pandas and Scikit-Learn
NER - Spacy
MERN Stack - MongoDB, Express Js, React Js, Node Js
PowerBI / Tableau for Visualizations

Technology Stack

10 of 13

10

Area of Expertise of Team Members

Member No.	Area of Expertise	Justification
1	Frontend - HTML, CSS, Javascript, React.js, Vue.js Embedded Softwares - Arduino Cybersecurity- Familiar with Basics and Foundations	Completed an internship on frontend and backend in Vue.js Did an internship as a software developer Completed an internship which had arduino Did courses on cybersecurity
2	Python : Beautiful soup used for crawler, Numpy, Pandas Backend : Worked with MongoDB Cybersecurity : Basic idea of TOR network.	Completed an in-house internship involving a scraper. Participated in hackathon and worked on backend.

11 of 13

11

Area of Expertise of Team Members

Member No.	Area of Expertise	Justification
3	Python - NLP, Cybersecurity - Pentesting, TOR browser, Web App Pentesting	Completed an in-house internship to develop a spam sms detection app, and external internship in web app pentesting
4	Frontend - HTML, CSS, Bootstrap, JavaScript, jQuery, React.js, Node.js Machine Learning - Matlab, Supervised and Unsupervised Learning, Neural Networks Embedded Softwares - Arduino, ATMega, STM development boards, Embedded C	Have completed an internship on frontend in React.js Completed a MOOC on Machine Learning hosted by Stanford University on Coursera. Have been in the college’s robotics team for 2 years as a member of embedded coding.

12 of 13

12

Area of Expertise of Team Members

Member No.	Area of Expertise	Justification
5	Fronted-Html,Css,Bootstrap,Javascript Data Science -Data preparation, Modeling,Evaluating model Performance on openSAP system . AI in Cybersecurity	Have completed udemy course on the topics . Have completed 6 week course on openSAP Training. Deloitte Bridge Session
6	Python - Data Science modules and Beautiful Soup for web crawler. Machine Learning - Basic understanding of various models Cybersecurity - Knowledge about foundations and TOR network. Development - Full Stack development using HTML,CSS, Django and complete MERN Stack. Data Visualization - PowerBi, Tableau	Worked on a Website Classifier using Machine Learning model as a part of an internship Pursuing interest in Cybersecurity, I have opted for Dept. Elective in same domain, as well as completed some Udemy courses to gain knowledge. Participated in a Hackathon to make a working web-application using the given tech stack and even worked on dashboards.

13 of 13

13

[1] UNODC World Drug Report 2020: https://www.unodc.org/documents/Focus/WDR20_Booklet_4_Darknet_web.pdf

[2] Understanding LSTM and its quick implementations: https://towardsdatascience.com/understanding-lstm-and-its-quick-implementation-in-keras-for-sentiment-analysis-af410fd85b47

References