Link to website Here
Tweet your comments/suggestion to @haveaseat_team5 [here] (https://www.twitter.com)
 Vimanyu Aggarwal |
 Sparsh Sidana |
 Amit Pandey |
 Jayam Malviya |
|---|
###Vimanyu Aggarwal(+1-669-254-7235) ###Sparsh Sidana (+1-669-265-5123) ###Amit Pandey (+1-669-265-7685) ###Jayam Malviya (+1-669-265-5125)
##Problem Statement: How many times has it occurred that you went to a diner and couldn’t find a seat? You probably ended up spending time in long queues, ruining your dining experience and getting frustrated – not knowing that there is a diner with available seats on the next street or there was a seat available just a few minutes ago when you stopped just for a small chat. Moreover, restaurant owners lose revenue (as well as food!) when a customer enters the diner right after the owner closed the cash counter machine. So, there is a need to develop a platform on which the consumers and the restaurant owners can communicate real-time.
##Target Audience: ###@Consumers: Foodies who wish to eat at a diner in a specific location without making prior reservations.
###@Owners: Restaurant owner who doesn’t want to waste food and have seats in their restaurant when nearby restaurant is full.
##Proposed Solution: ###@Consumers: Foodies can now look up in the app and see if there is a seat available at the location where they want to eat. They can submit the request to book the seat, and the app will give them a time-limit to reach to the restaurant. If they reach the restaurant in the specified time, the owner can check-in them and mark their seat as booked. As soon as the owner marks the seat as booked, it will be reflected across all the apps in real time.
##User Architectural Flow Diagram
###@Owners: Owners will get a notification via ML algorithm one-hour prior to closing time to send tweets/messages to the foodies based on the historical data accumulated by the app. Owner can then decide whether to call it a day & go home or to send the tweets to the foodie popped-out by the ML algorithm. Owners also get to disapprove the request for booking in case a favorite customer comes in before the ones who booked the seat via the app.
###Owner Architectural Flow Diagram
###Welcome Page for Customer and Owner
This is the default landing page when a user (either customer/owner) enters the website. This page contains the basic introduction to the concept of “Have A Seat” as well as provides the links to facebook page of Have A Seat, contact information form at the end. User input fields are protected from XSS attacks and SQL Injection.
###Search page with search results and the “available seats“ on the restaurants
This page gives the live status of seats availability when user searches based on location and/or restaurant name. User input is filtered in the search box so that the search works even if the user enters keywords without spaces, in different cases (upper/lower) etc. Customer need not login to view the live seat availability.
###Seats view of restaurant for customer persona :
Customer can select the available/white seat(s) from the seat layout. User is given a ticker to contact the Have A Seat development team in case he/she has any concerns. Customer can also send reviews regarding the restaurant on this page. The user reviews are fed to a sentimental analysis using NLTK (Natural Language ToolKit) library available in Python-Flask framework. This analysis gives the insights in graphical format to the restaurant owner on his/her dashboard. Customer can only book seats after logging in but can view available seats without login.
###User booking seats layout at the restaurant:
The customer can book seats on this page by clicking on the seat he/she likes based on their positions as per the layout. The customer can book multiple seats for dining. The color of seat changes to Green when the customer selects the seat. The seats selected are displayed below the checkout button whenever the customer selects one and the list of the seats is updated dynamically with the customer’s selection.
###Customer un-booking from the selected seats
The customer can unselect the seats he/she had selected earlier by clicking on the unbook book button corresponding to the table number which is displayed below the the checkout button on the page.
###Customer trying checkout without logging in
As we can observe the application allows the user to browse through the application, and view the currently available tables at the restaurants, but in order to check out his table he must be logged in. This screen depicts Login message that the user is given when tries to book a table without being logged in.
###Login Page
User must use this page in order to login to the application, the same page is also used by the owner to login. Upon successful authentication the user is redirected to the home page. And if it’s an owner he is redirected towards his dashboard.
###Encryption in Database
As can be seen the passwords of the users are encrypted using UTF-8 128 bit encryption before storing in the database hosted on cloud using Python’s bcrypt library.
###Home Page showing Customer login
User login can be verified from the header. Customer is greeted with a Hi message in the header followed by his name. It is only after the successful login of the user, it is possible for him to book a table.
###Customer successfully booking the Tables
This message depicts that users booking has been successful and he is now ready to dine at his chosen tables at the restaurant. The message also tells how many tables were booked in his checkout. The status of the seats selected by him change from “Available” (White) to “Orange” (Booked). This change is reflected across all displays in real-time. Customer also gets an e-mail confirming the booking. After this page he is redirected to the Thankyou page.
###Thank You Page for customer with timer
This is the Thankyou page to which the user is redirected after his successful checkout of tables. He is also given a timer at this page indicating the time left for his booking to be valid. After the timer expires the user's booking is no more valid.
###Owner Dashboard View of Restaurant
As he logs as the owner of the restaurant,he is redirected to the restaurant he owns, this is the home dashboard of the owner, where he can see the current status of all the tables in the restaurant.
###Features available to owner Owner can see the status of seats in his own restaurant. He also has the flexibility to change the status of the seats as per his will.
###Owner making the changes in seat availability status Owner can change the state of the table at the restaurant and it is reflected in the system instantly. He can make any table available, Booked or Dining as he feels appropriate. These changed state of tables will be reflected to the user in real time.
An Amazon Echo device will be deployed at the restaurant which will be helping the owner with confirming booked seats. Alexa will be assisting the owner by confirming customer details when they reach the restaurant after booking seats using the Have A Seat application. Alexa just need customer’s name and the table number they booked and Alexa will confirm the booking if the details presented by customer are valid.
###Tweet Functionality
Owner can tweet about the instant offer that he wants to give to its customers. Tweet can be send using haeaseat_team5 twitter handle. Customers need to follow only tweets at @haveaseat_team5 about the latest offer that the restaurants affiliated with us offer.
###Targeted Marketing
As marketing strategy can be implemented in two ways - Give huge discounts to the low frequency customers (Exploration Mode) or give moderate discounts to high frequency customers (Exploitation Mode). Epsilon-Greedy Algorithm is a simplistic algorithm that came out as the direct result of Multi-Armed Bandit Problem posed by Microsoft Research. It has been proven that this method beats A/B Testing everytime.
Emails crunched out by Epsilon-Greedy Algorithm enables the owner to do targeted marketing. This is further refined by providing time-slots to the owner which specify in which time slot does the owner wants to do targeted marketing. Because the algorithm plays on large data set - the output is formed by a cron job which runs on BlueMix Platform at mid-night and is stored on Flask Server. This saves the owner from waiting while the algorithm is crunching out the emails.
###Sentimental Analysis on user reviews
Owner can click on the Get Graph button to get the comparative analysis of the reviews that the customers have submitted on the Have A Seat website. The Have A Seat team provides this graphical view to the owner by performing sentimental analysis using NLTK (Natural Language ToolKit) library on the reviews submitted by the customers.
###Contact to Have A Seat Team
Owner can send email to Have A Seat team, as per his requirements & needs. And team will be available to resolve their issues. This functionality can also be used to share instant feedback between Owner & Have A Seat.
##Future Scope: Wider ML integration to suggest appropriate restaurants to the foodies based on their eating habits, and to suggest owners what food to prepare and how much in case a customer has booked a seat. With much data (real and accurate), we can scale the app to release APIs which can be used to develop other apps (or can also expand this app) and give recommendations of available seat based on time/season to the users.
##Benefits: • Increased revenue of the restaurant owners by allowing them to contact prospective customers thereby allowing them to fill up the seats at the last moment. Improve the opportunity costs for the restaurant owners when they have seats available. Enables the owners to do targeted marketing and gain automated insights from the user reviews in graphical format. Owners can reach out the customers which were previously out-of-scope by using Have A Seat Database. • Lots and lots of happiness to the Foodies!
##Overall Architectural Flow Diagram:
