According to recent studies, 1 In 10 Indian Adolescents Faces Cyber Bullying, half don’t report [1]. Another survey conducted by National Family Health Survey (NFHS) conducted in 2015-16 and compiled by NCRB shows that estimated 99.1% of sexual violence cases are not reported [2]. And there are many other surveys about unreported abuse cases. The common reason found behind all these unreported cases is the hesitation of the victim to open up to others for asking help which is further strengthened due to personal reasons.
Project Saathi aims to promote justice for all who have gone through harassment, bullying or violence in any form and shape. Saathi, the chatbot uses machine learning as a tool for faster and effective actions.

In institutions of higher education (IHEs), 23.1% of female and 5.4% of male undergraduate students experience rape or sexual assault by use of physical force, violence, or incapacitation [3], yet an estimated 80% of incidents are not reported [4]. Rape or sexual assault are about half as likely to be reported to police as robbery (54%) and aggravated assault (58%), with the former being reported in only about a quarter of all cases (23%) [5]. Social stigma surrounding sexual crimes likely contributes to this low level of reporting

Our objective is to provide help from legal authorities to those victims who do not even report their cases due to reasons such as lack of sources, neglect, fear of privacy invasion and many others. Saathi is an exceptional chatbot which helps the user to muster up the courage to open about the unpleasant experience they have gone through, and encourage them to report and get help from legal authorities or NGOs. It provides a fast response to the users, at the same time, also maintains their privacy. Even if the user prefers to not report, they can at least share their pain which they usually wouldn’t have.

Lending a helping hand to all our users and simultaneously safeguarding their privacy is one of the main features of the project. This is managed in following steps: Step1: A unique 16 digit code will be provided to each user which will be required to copy for later use. Step2: Consent of the user will be taken to pass on the matter to legal authorities or NGOs. Step3: If the consent is given, Saathi, the Chatbot will ask for the email id and 16 digit code of the user. Step4: Only the email id will be forwarded directly to legal authorities to initiate contact with the user. The User needs to provide a 16 digit code to confirm the final ascent.

Saathi is unparalleled and useful and the features which advantage the innovation that lies within the core of project are:

• Privacy at its core
• Free of charge
• Fast and Quick with response
• Promotes justice and peace
• Promotes gender equality
• Reduces cyber bullying
• Evaluates and categorize user’s ordeal (planned, to be introduced in future release)

Google:

• Firebase (Firestore, Hosting)
• Progressive Web Apps (PWA)
• TensorFlow JS (Work undergoing, will be published in next release.)

Open Source:

• NodeJS
• NLP.JS

You can visit the Live Demo to view the app directly without any need of installation.
If you are working on a local workstation follow the instructions to use this app.

• Method 1: Downloading the ZIP from GitHub web.
  Step 1. Download the ZIP archive of this repo on your local computer.
  Step 2. On your local computer, unzip the archive.

• Method 2: Using 'git' to clone the repo.
  Step 1. Install git on your pc if not already (Here's a help from GitHub).
  Step 2: Run git clone https://github.com/ayushanand18/project-saathi\

• Next Steps
  Step 3: Navigate to the root directory of the project (on your local system).
  Step 4: Navigate to the public folder.
  Step 5: Open index.html on your browser.

The PWA is live on https://project-saathi-43be5.web.app/, Thanks to Firebase Hosting and Plasmic!

We are open to collaborating with interested NGOs working to eliminate gender-based crimes.

Ayush Anand (Development)
Nishtha Bhardwaj (Testing and Documentation)

We are working on some possible enhancements but they are taking more time than the hackathon.\

• Named Entity Recognition from descriptions of victims e.g. place of assault, age, demographics, etc to better help policing and policy making.
• Visualization of insights from the decriptions of victims.
• Categorization of abuse types using ML (TensorFlow).
• Enabling Live human agent support (on request).

1. https://www.indiaspend.com/1-in-10-indian-adolescents-faces-cyberbullying-half-dont-report-study/
2. https://www.livemint.com/Politics/AV3sIKoEBAGZozALMX8THK/99-cases-of-sexual-assaults-go-unreported-govt-data-shows/
3. D. Cantor, B. Fisher, S. H. Chibnall, R. Townsend, H. Lee, G. Thomas, C. Bruce, and I. Westat, “Report on the aau campus climate survey on sexual assault and sexual misconduct,” 2015.
4. N. R. Council, Estimating the Incidence of Rape and Sexual Assault, C. Kruttschnitt, W. D. Kalsbeek, and C. C. House, Eds. Washington, DC: The National Academies Press, 2014. [Online]. Available: https://www.nap.edu/catalog/18605/ estimating-the-incidence-of-rape-and-sexual-assault
5. B. of Justice Statistics., Criminal Victimization, 2016., 2017. [Online]. Available: https://www.bjs.gov/content/pub/pdf/cv16 sum.pdf