Howdinn


Discover a new level of self-awareness and emotional understanding with Howdinn. Gain valuable insights, receive personalized advice, and embark on a journey towards improved emotional well-being.

How are you doin? Howdinn is the service that allows users to record videos on the website, analyze their emotions using the recorded video and voice, and provide emotion analysis charts and considerate advice and recommendations.

Problem Statement


  1. Privacy Concerns: People who write diaries may desire to keep their thoughts and emotions private and not share them with others. However, they still yearn for a way to receive feedback or advice without feeling embarrassed or exposed.

  2. Lack of Objectivity: Individuals often find it challenging to objectively analyze their feelings and emotions solely through written diaries or self-reflection. They may struggle to accurately interpret and understand their emotions, hindering their ability to gain valuable insights and make meaningful improvements.

  3. Emotional Expression: Written diaries may fail to capture the full range and depth of emotions due to subconscious filtering. People tend to censor or modify their true emotions when they write, resulting in a less expressive and potentially less useful diary. To facilitate emotion analysis and provide effective advice, incorporating visual and auditory information is seen as more valuable.

  4. Need for Text Diary Alternatives: There is a perceived gap in the availability of alternative diary formats that can better capture and convey emotions. Traditional written diaries may not sufficiently cater to the emotional needs of individuals seeking objective analysis and guidance. Exploring options beyond text-based diaries could offer enhanced emotional insights and more helpful advice.

What it does


  • Questionnaire-Based Sentiment Analysis: Howdinn presents users with various thought-provoking questions about life, allowing them to record video responses. The service then utilizes advanced sentiment analysis techniques to analyze the emotions expressed in the videos, providing a comprehensive understanding of the user's emotional state.

  • Emotion Visualization: Howdinn visually presents the most frequent emotions experienced by the user through intuitive charts and graphs. This visualization helps users recognize patterns, identify their dominant emotions, and gain a clearer understanding of their emotional well-being.

  • Personalized Advice: Based on the emotion analysis results, Howdinn offers considerate and tailored advice. This advice aims to provide guidance and recommendations for managing emotions, improving overall well-being, and addressing specific emotional challenges or concerns.

  • Recommendations: Howdinn goes beyond emotional analysis and advice by offering personalized recommendations for movies, food, and activities. Taking into account the user's emotional state, the service suggests movies to watch, food to indulge in, and activities to engage in, all geared towards enhancing their emotional well-being and enjoyment.

Howdinn is a comprehensive web service that enables users to record video responses to thought-provoking questions about life. By analyzing the sentiments expressed in the videos, the service provides users with a visualization of their most frequent emotions, along with personalized advice to support their emotional well-being. Additionally, Howdinn goes the extra mile by offering recommendations for movies, food, and activities that align with the user's emotional state, fostering a holistic approach to emotional understanding and improvement.

How we built it


  • AI Integration: Our team integrated the Humm AI model and GPT-4 from OpenAI. Humm AI analyzed the video and voice recordings to identify the top 5 frequent emotions present in the user's video, along with corresponding sentiment scores. This information served as input for GPT-4, which generated the most suitable and useful advice and recommendations for the user based on the emotion analysis results.

  • Backend Development: Flask, a Python web framework, was used to build the backend infrastructure of Howdinn. It handled the communication between the frontend and the server, processed user requests, and managed login sessions for future features. Also, all the AI APIs were built in Flask. Additionally, the backend was responsible for storing and retrieving the final reports and providing statistical data.

  • Frontend Development: The frontend of Howdinn was developed using HTML, CSS, and JavaScript. These technologies allowed for the creation of a user-friendly interface where users could record and submit their video responses. Google Chart API was utilized to visualize the most frequent emotions with intuitive charts and graphs, enhancing the user's understanding of their emotional patterns.

  • Server and Database: Google Firebase was utilized as the server and database for Howdinn. It provided a scalable and reliable infrastructure to store and manage user data, including video recordings, emotion analysis results, and final reports. Firebase's authentication services were used to secure user login sessions.

By combining AI technologies, a robust backend infrastructure, an interactive frontend, and secure cloud services, the Howdinn team created a comprehensive web service that empowers users to analyze their emotions and receive personalized advice and recommendations for improved emotional well-being.

Challenges we ran into


  • Integrating Sentiment Analysis Model in the Web: Integrating the sentiment analysis model into the web application posed technical challenges. Ensuring seamless communication between the frontend, backend, and AI model required careful design and implementation to achieve accurate emotion analysis results in real-time.

  • Generating Better Results in GPT: Improving the quality of advice and recommendations generated by GPT presented a challenge. We did a lot of experiments to figure out the best prompt that generated the appropriate results.

Accomplishments that we're proud of


  • Integrating Hume AI Streaming API in our Flask server
  • Tweaking the Prompt of GPT4 and got better results
  • Completed the project within a time
  • Combined inference model and generative model together and yield valuable service

What we learned


  • Better and border understand of LLM and Inference model
  • How to use various AI API in web
  • How to run and deploy server with Flask
  • Using Hume’s Streaming API to figure out general mood of the person

What's next for


  • Annual Mood-Changing Statistics: Howdinn aims to provide users with an annual mood-changing statistic feature. This would enable users to track and visualize their emotional patterns and changes over time, offering valuable insights into their long-term emotional well-being. The feature could include charts, graphs, and personalized summaries to help users better understand their emotional journey.

  • Speech-to-Text (STT) Integration: To enhance the user experience, Howdinn plans to integrate a Speech-to-Text (STT) model. This integration would allow users to convert their spoken responses into text, which can then be utilized as input for the GPT model. By incorporating STT, users can have more flexibility in how they interact with the service, providing a seamless and natural conversation experience.

  • User Persona Creation: Howdinn aims to leverage the input data from users to create personalized user personas. These personas would be based on the user's expressed emotions, interests, and preferences. By generating a user persona, Howdinn can offer more tailored and relevant advice, creating a more engaging and personalized interaction where users can have conversations with their own personas to gain additional insights and support.

  • These future plans reflect the commitment of Howdinn to continually enhance the service and provide users with innovative features that promote emotional understanding, self-reflection, and personal growth. By expanding into annual statistics, integrating STT capabilities, and introducing user personas, Howdinn seeks to create a more comprehensive and immersive user experience, fostering deeper emotional connections and insights.

Why Howdinn?


Self-Reflection and Emotional Understanding: Howdinn allows individuals to engage in self-reflection and gain a deeper understanding of their emotions. By recording video responses to thought-provoking questions, users can express and explore their feelings in a more comprehensive way. This process can help individuals become more aware of their emotions, identify patterns, and gain insights into their emotional well-being.

Objective Emotion Analysis: Howdinn employs sentiment analysis techniques to objectively analyze the emotions expressed in the recorded videos. This provides users with a data-driven perspective on their emotional state, helping them see their emotions from an outside perspective and potentially uncovering patterns and insights they may have missed through self-reflection alone.

Personalized Advice and Recommendations: Based on the emotion analysis results, Howdinn provides personalized advice and recommendations. This guidance takes into account the user's specific emotional needs, helping them navigate challenging situations, manage stress, improve mood, or enhance overall well-being. The service aims to provide practical and considerate suggestions tailored to each individual.

Emotional Support and Validation: Sometimes, individuals may hesitate to seek feedback or advice from others due to embarrassment or privacy concerns. Howdinn offers a safe and confidential space where users can freely express their emotions and receive support without fear of judgment. This can be particularly helpful for those who may not have immediate access to a support network or prefer a more private and introspective approach.

Enriched Emotional Experience: In addition to emotion analysis and advice, Howdinn goes beyond by providing recommendations for movies, food, and activities based on the user's emotions. This adds an element of enjoyment and exploration, helping individuals find ways to uplift their mood, engage in self-care, and enhance their overall emotional well-being.

Built With

Share this project:

Updates