ML APP | FISH SPECIES PREDICTION

In this article we are going to visualize and predict the fish species using various illustrations and python libraries. We have downloaded  the dataset for this particular web app  from kaggle.com. This web app is built with python streamlit library. We have also included exploratory data analysis and data visualization.

The dataset has seven columns namely Species, Weight, Length1, Length2, Lenght3, Height, Width of the fish. In this Machine learning web app we are predicting the fish Species based on other six attributes. This dataset include seven most common fish species namely Bream, Perch, Pike, Whitefish, Roach, Smelt, Parkki. We have trained our machine learning model using this data to predict the fish species on the new data. On the home page of the app user needs to enter all the input attributes and click on predict button to find out the fish species. The result obtained is the image of fish species and the type of species.

 User can navigate to Home, EDA or ABOUT section rom the navigation side bar. The EDA section has EXPLORATORY DATA ANALYSIS and DATA VISUALIZATION sub-sections. Clicking on Show dataset checkbox shows the dataset used for this web app. Also we have show columns , Show shape, Show summery, and correlation check box. 

Libraries used for data visualization are matplotlib, seaborn and PIL for displaying images in the app.

The DATA VISUALIZATION section has simple value plot, scatter plot and box plot checkboxes. Simple value plot gives count of each species in the dataset. We have used scatterplot for this dataset to check for any outliers in the dataset. Analyzing the outliers we have filtered out the outliers to increase the performance of our model.  You can also see the boxplot for the species column by clicking on Box Plot in the app.

The details about this app and how to use it, is given in the ABOUT section. The brief about the input attributes is also provided in this section.