Skip to content

Proposal

Jump to bottom
JaehoNam edited this page Oct 5, 2019 · 1 revision

Proposal

Overall Description

Photo based, personal experience-based menu, restaurant, recipe recommendation and sharing platform

Target Customers

  1. People who think a lot about what to eat

  2. People who lean on other’s experience, not him/herself when they find what to eat

  3. People who take a picture about food a lot but don’t use them

  4. People who loves to cook, but finds it hard to find a good recipe for them

Motivation and Abstract

 The experience about food is very personal, but people tends to lean on other’s experience, such as power bloggers, Instagram, or MangoPlate when they choose what to eat, what restaurant they’ll go, or what recipe they will use. In addition, many people tend to take a picture of food they are eating, but it is just stored in the gallery, neither classified nor used. People eat about 1200 meal a year, assuming three meals daily. In 10 year, they will eat about 12000 foods. These are relatively abundant data since rarely do heavy bloggers post over 10,000 articles. In addition, since nobody has exactly same taste, no other reviews are more valuable for themselves than their review and picture of food they eat.

 Therefore, we will use this neglected data to personalize their experience about foods and share it with others. People who take food pictures don’t write reviews because of cumbersome categorization, high entry barriers, and no utility of review itself. If someone leave a simple line of feeling along with a picture of the food, our web-service will automatically analyze and categorize it. After then, when in need of information about foods, we will suggest information which based on everyone’s own experience.

 When customers want to eat a particular menu while looking at a picture, for example, pizza, our web service will recommend a nearby pizza restaurant, but analyze the information they've eaten so far, giving priority to the side that's well-reviewed. If they want to make themselves a meal, it will automatically bring them the entire recipe so that they can purchase the all the ingredients they need for the recipe right away. When they are unsure what menu they will eat, it will recommend something in the larger category (ex: Korean food, Chinese food, etc.) or something they have eaten frequently recently, or something they used to eat frequently. In addition, when they want to try new foods or restaurants, it doesn’t just recommend the ones with high ratings, but offer better experience optimized to individuals as much as possible, giving the priority to the restaurants and foods which is highly rated within people who have similar tastes with them.

Instructions

  1. Main menu consists of upload, my gallery, friends’ gallery.

  2. When user uploads photo, get relevant information including its name, location, one-line review, and rating with it. By default, location information could be obtained from photo’s metadata, leave one-line review as an option.

  3. My gallery is ordered by day like calendar by default. It can also be ordered by category(ML), menu, place, one-line review, star rating(ML).

  4. When user clicks photo in the gallery, show up menu’s name, location, one-line review, and rating. And additionally, shows up option to “A) order this food B) cook this food C) recommend this food to friends”.

    1. When user orders food, provide recommended restaurant information based on current location information. However, if user has record of visiting some of restaurants, take his preference into account while sorting restaurants.
    2. When user wants to cook food him/herself, obtain recipe from external recipe platform. Provide “cart” system to buy ingredients for recipe at once, and if some of the ingredients couldn’t be purchased, redirect them to external website.
    3. When recommendation is provided to other friends, consider their preference. In other words, provide suggestion to friends who are likely to like the food too, not user’s close friends.

  5. Menu suggestion is provided in my gallery. Suggestion is based on category, recent history, old history, one-line review, food that many friends ate.

  6. Account system should be used. User can add friends by linking to their SNS account, or from their contacts.

Essential Features

Every feature related to personalization. Upload, edit, deletion should be available. Gallery’s sorting system, ordering, cooking, and menu recommendation system.

Non-essential Features

  • Feature to interact with friends added on account system.
  • Reply on or bookmark friends’ gallery using account system.
  • Link google or SNS account to find friends or friends’ friends.
  • Link google drive to upload all of photos they’ve taken automatically.

Demo and Test

 Our services should be tested mainly if it provides proper personalization, and menu sorting / one-line review analysis based on ML. Also, when menu is selected, it should be checked if it provides correct restaurant and recipe. Suggestions should be personalized enough for users to feel that they’re based on their own experience.

  1. Many people already have many food photos, so upload them and create many accounts.

  2. In one of the accounts, test sorting, suggestion, and recipe search feature based on already uploaded database.

  3. Upload many photo of specific menu and test if sorting and recommendation still works properly.

  4. See if other account’s gallery is visible, and check if feature of finding other accounts with similar preference works well.

Clone this wiki locally