sntyagi1 - Tumblr blog

[sntyagi1]

TOOLS USED IN THE PROJECT

1. MYSQL workbench

2. Python- Jupyter notebook

3. Microsoft Excel

4. Tableau

[sntyagi1]

Data Analysis

INTRODUCTION

Data Analytics is a very vast field in which many operations reside which needs to be done before preparing the final conclusion for the client in regards to the problem that they may be facing. There are many steps that needs to be taken before completing the presentation that needs to be shown to the client, so that they can make correct decisions for their organizations.

 There are variety of tools which are required to complete the whole process of the data analysis of the large amount of data given to the analyst. Tools such as Microsoft Excel which is an excellent example in which we can prepare any type of charts by modifying the default chart types presented to us in the tool itself. We can perform different types of mathematical operations on the given data to fully understand the data before performing any prediction analysis, so that we do not make any mistake further in the future process.It is an excellent tool to visualize the data.

PROBLEM STATEMENT

The main objective of the project is to provide the solution to the problem faced by the client, using the data provided from them about their organization and give them the result taken out using complicated mathematical procedures, so that they can make appropriate decisions which will favor them.

SOLUTION

The data is taken out from the server by the clients and it needs to be taken out from there. We only use the select statement on the data and not change the underlying structure of the data.The data is uploaded on server because the once the client sends data to an organisation for data analysis, it is not necessary that only one person is working on that data, but a bunch of teams simultaneously may be working on different aspects of the project using the same data.

Sql helps in understanding the data and taking out some relevant information before performing any operation on it. This information helps to shape the structure the result and will tell us about the different percentages involved in the makeup of whole orders. One of the client needed information about the orders of mobile manufacturers and so we had to use sub queries to take out the different percentage from the different companies using the mathematical tools provided on the sql itself.

 Data analytics is about getting some useful information out of the raw data which companies collect by selling their resources or products. Sometime they need to know more about their own products, they want to find out which product is selling better and in which region. Different products sells better at different time and under different conditions. For example sweatshirts are sold in large quantities in winters and wife beaters in summers. Some products needs to be changed according to the regions also, like models of same car may differ in some slight way for different country. All of these things can not be determined by the organization itself, so they need data analysts, who uses tools like tableau or power bi. Tableau helps in preparing charts and tables to show different compositions of the data in a quick and easy way. Using this data we can come to some conclusion about the products or services which the client is providing and help them to understand whether they need to change in some ways or they some keep doing what they are doing.

LEARNING OUTCOME

After completing the project, we can understand that after getting the data from the client and getting started with the prediction is not the right way to proceed, first we have to fully understand the data and get complete understanding of what the client’s need is. This can be done by making charts and tables or by whatever ways we need, to fully understand the data. After working on this project I have understood that planning s very important and the needs which the client has should be very clear before performing any steps in the direction of the solution.

EXPERIENCE

Working in an office where there are deadlines which will not get extended, was a difficult and challenging task. There was a few weeks of time where I was taught about the work that I would be doing. My mentor at the organization were very patient and helpful, whenever I had any doubt they would explain the topic again till I was clear, in their free time of course because work come first.

[sntyagi1]

USER INTERFACE 

[sntyagi1]

MESS IN BENNETT

In many organizations whether educational or industrial there are group which caters the people of the said organization and have some kind of process to help them. This application will help all these organizations to make their work easy and complete the work in a very ordered manner. As the basis for all the process is same so the application be change with accordance to the users wish. As the application is simple to use, it can be operated by anyone carrying an android device and connected to the organization database to login to their own account to start using the application.

TIMELINE

    Week 1

The first week was spent just accumulating back with the environment of the university and not much was done in terms of project development. In the first week the focus was mainly on the introduction with the new subjects given or chosen by us for the new semester. As we returned from our internship, we had to submit a full-fledged report about the same and a video was to be uploaded on our YouTube channel and all of this took my whole time for that first week back at college. Also, as it was the first week no one was in the mood to start brainstorming for the idea of the project so early.

      Week 2

The second was spent brainstorming for the idea for the project. As I did web-development in my internship I wanted incorporate that learning in my project. So, went around the teacher’s department and asked for their help in choosing a project which will not only be impressive but will be completed in the given amount of time. But that was of no help, I was not able to get any good ideas from the CS department and to rely on the internet only. I knew that I wanted to combine web development and machine leaning in my project, but I didn’t know in what way.

    Week 3

In the third week I started researching on the internet for the ideas for the capstone project, this again proved to be quite difficult because some projects which I found were of very low level and some were of so high that I was quite shocked that people had done sch impressive work in their B.Tech degree only. The idea of my project from my friend/roommate, he told me about the problem faced by the mess committee and how much food is wasted when some students decide to skip their meals, particularly on weekends. All of this resulted in the idea for my project.

    Week 4

In the fourth week I started focusing on the methodology for my project. It was a difficult process as I must start the course on android development, because before this I have not dwelled in the development of smartphones applications. This was challenging as I had to start writing for the first milestone also. The first milestone was a big document which ate up a lot of my time. Finally, the first milestone was completed, and I started working for the development of my project with focus on the development pf the machine learning model and taking out the data.

  Week 5

After submitting the first milestone, the placement preparedness classes started and the companies started coming, which ate up a lot of time. As I had to prepare for the aptitude tests and prepare myself for the interview process, all of my time was spent focusing on readying myself for the placement drive. But in between all these procedures I still was able to focus on completing the android course(basic) on Udacity. I also started clicking pictures o the dishes served to us in the mess and collected images from the internet of the three dishes which I focused on first.

    Week 6

In the third week the companies started coming and each day at least one company was in the campus. The progress of the development of the project was slow and I basically did very little in this week. I started working on the model to recognize different dishes. Last semester I learned deep learning in which I learned CNN, which was a method to recognize images.

I started working on the development for the model to differentiate between different food items, but to train the model I need a lot of images and a variety of it with some data specially for validation.

    Week 7

In the seventh week I started working on the user interface of the app and started making the registration and login page for the application. I used online available firebase database available to us free from google. I connected it with my android studio and used to allow access to the users. I stated training my model VGG16 but it is taking a lot of time to train the data on azure notebook and also some considerable amount of time is spent on the validation part. I was able to do this much only as I was busy in the placement process.

    Week 8

In the eighth week I stated writing the second milestone, which again is a big document and is taking a lot of time to complete the document. This week is especially hectic because of the coming mid-term and the placement drive which is happening. As I have not been able to secure a job yet my main focus is on the placement process, the only I will be able to spend all my time on the development of the project. The development of the project again will be slow in the next week also because of the coming test and the lab evaluation in the subject of RPA.

Week 9

In the coming 9th week I will be focusing on the mid-terms and the evaluation of the second milestone. I will also try to increase the speed of my VGG16 model and try to bring some changes in it so that the latency will be reduced. Also, I will try to add some basic functionalities in the application I am developing at the same time with the model. If by then I am not placed I will be revising all my previous subjects to improve my chances in the next upcoming placement drive. But the focus will jot be shifted from the project.

 Week 10

In the 10th week I will focus on the model to predict the favorite dishes of the users. That will need some data to train but as I will not have data from the users, I will try to make up some data and feed it to make it work. At the same time, I will probably research how to upload my models on the server so that the android application can communicate with model to predict the results. As this will be new for me so it will probably take up a lot of time, but the necessary work will have to be completed.

  Week 11

 As the deadline for the project submission will be coming close, I will have to start working very quickly and start focusing on the app development, in a fashion that the users will be able to understand how to use without any instructions given to them. The user interface should be very easy for any user to understand and, I will focus on how to reduce the latency of getting the output from the server. If there are any documents to be written I will also have to spend time on them. But at tis time I hope to complete my project.

  Week 12

At this point of time I will focus on finetuning any development left in my project. I will probably try to improve my application and try to train my model as much as possible in the time left. I will probably try to improve all the aspects of my project and if there is anything that I may have missed in the development procedure will try to remove it and finish my project on time. Also, will have to test my application in real time and will have to receive the feedback from the users, which again is an essential part of any product development.

DESCRIPTION

The project is using machine learning in an android application which is not very easy to achieve. The application will be opened by the user and they will be met with the registration tab in the starting which is connected to free to use database “Firebase" available to the people on the net, this database is generally used for the mobile applications only. In this the data which is being sent from the application is sent will be saved. Also, the data can be read from here if there comes, if any need arise to use some kind of functionalities which require to take out data from the database, it will not be difficult.

The user will have to submit their email id and must make a password to register in the application and to start using its functionalities and same goes for login also. After logging inn, the user can select the menu button to see the menu for that day. Also, there will be button identify the different dishes available in the menu.

This will automatically open the camera of your phone for you and you can click the image of the food. As the model is using a couple thousand image to identify the dish, the captured picture should be clear and not blurry and the focus should be on the actual dish and not other items, then only the model will work properly. There are thousands of images that are used to train this model and some few to test and validate the model. Also the process to convert the python file into the file which can be used in the android project is also difficult.