Estimate Fish Population and Price

Using Census Gov Fishery Data

Joseph I
--

Introduction

US Census Bureau: Fisheries, Aquaculture provides public access to fish population and price in the United States for the past 25 years.

By using this historical data, we can build a simple forecast model to predict each fish population and their prices.

Preparation

The data was downloaded from census.gov, under Forestry, Fishing, and Mining section
- 897 - Domestic Fish and Shellfish Catch, Value and Price by Species

There are 36 different fish species tracked in the data set. Out of the 36 species, 9 of them are shellfish, which we will skip for this exercise.

The data set contains each fish's

  • population,
  • value (in USD), and
  • price per pound (in USD) from 1985 to 2009

Analysis

After simple data cleaning, forecast modeling is done using expotential time-series formula.

  • For example: to find forecast for "Anchovies"
library(forecast)
anchovies_time_series <- ts(as.numeric(df[df$Species=="Anchovies",2:26]), frequency=1)
anchovies_fit <- ets(anchovies_time_series)
anchovies_forecast <- forecast(anchovies_fit)

Result

This is the forecast result for Anchovie

plot of chunk unnamed-chunk-3