DataFirst

Inspiration

Seeing an increase in threats for user data and privacy both personally and with friends/family motivated me to creating a user-friendly solution. This project would help innocent people suffering from the consequences of people with malicious intent.

What it does

Data first is an intuitive app that can flag suspicious calls, text messages, and emails a user may receive. The app uses existing data to determine if a user's personal information may be stolen. This includes banking details, SSN, personal details(full name, date of birth, residence), and even scammer like speech. After detecting suspicious activity the app provides several notifications based on the threat level. There is even a real time threat meter if a user wants to detect scammers on a phone call. Furthermore, an integrated chatbot also allows the user to screenshot any suspicious text messages or emails, with the bot responding with a threat score as well as next steps.

How I built it

DataFirst was built for iOS using Swift, SwiftUI, on Xcode. It uses native Apple frameworks including Core ML, SoundAnalysis, AVFoundation, Speech, and Vision to analyze voices, suspicious language, and screenshots directly on the device. Local notifications, and a privacy-focused chatbot provide real-time warnings against scams and also provide next steps to secure a user's sensitive data.

Challenges I ran into

A challenge I ran into was making sure the app was user-friendly, despite it being a technically focused app. Ensuring it remained simple and intuitive was crucial so that users of all technical abilities could use it flawlessly. I ensured to spend a lot of time developing the UI so even users that are older can use the app without assistance to detect scams. Another issue was ensuring user privacy, as DataFirst needs access to one's microphone to provide real time warnings. It wasn't necessary to have access to a user's microphone all the time, which would violate their privacy. Instead, I implemented a simple toggle within the app, this way a user can manually turn it on when getting a suspicious call.

Accomplishments that I'm proud of

I'm proud to have put my iOS development skills and built a practical application that users of all technical abilities can use easily. Building a complete prototype of an iOS app that can provide real-time warnings, screen audio for suspicious activity, and an integrated chatbot are all features that I'm proud of. Furthermore, the idea to protect the everyday user and ensure that their sensitive data is protected is one that pleases me.

What I learned

I learned that detecting scams requires more than identifying suspicious words because context, urgency, voice characteristics, and user behavior all matter. We also gained experience with Apple’s on-device frameworks, real-time audio processing, privacy restrictions, and designing alerts that are helpful without causing unnecessary fear.

What's next for DataFirst

Next, DataFirst plan's to train and validate a dedicated Core ML deepfake-voice detector, so that users can know if a video or audio is AI generated. Supporting additional languages and testing DataFirst with seniors, families, financial institutions, and communities to get real world feedback.