Facebook Audience Insight on Food Choice

Published: 16 November 2020| Version 2 | DOI: 10.17632/9dbd9jcdvs.2
Contributors:
Lidia Mayangsari,
Muhammad Azizul Hakim

Description

We collected the data through an online platform by performing a query search on Facebook Audience Insights Interests. The keywords that we use in the question quest are based on the United Nations Food and Agriculture Organisation (FAO) Food Balance Sheet (FBS) which is retrieved from FAOStat in May 2020. The data was gathered between 15 May and 2 July 2020. With a sample size of 100-150 million viewers or about 36.95 per cent-55.43 percent of Indonesia's 2019 population, we limited our sample to Indonesia. The dataset is made up of ten tables that can be separately analyzed. For each table, we carry out exploratory data analysis (EDA) to provide more insights.

Files

Steps to reproduce

Data Retrieval Facebook Audience Insights: how to retrieve data is in the Specification Table and how data were acquired is by running query search on Facebook Audience Insights' Interests. The full tutorial is here: https://www.facebook.com/business/news/audience-insights. If the link above cannot be opened, it means you need to create an ads account, or a business manager first, the complete guide is here: https://www.facebook.com/business/learn/how-business-manager-works/guide The data retrieval for FAOStat can be studied here: http://www.fao.org/faostat/en/#data/FBS. Specification table: A description of the data collected, following the DIB template guidelines. Table 1: Description of the data collected, following the DIB template rules. List all table names and their descriptions, and the variables for each table Table 2: Supplementary data descriptions Table 3: Output statistical tests using ANOVA. If the P-value <= 0.05 and F> Fcrit, then the null hypothesis is rejected, or in other words, statistically significant, there are one or more variables that have an influence on the Domestic Supply Quantity of foods in Indonesia. Table 4-7: Pearson correlation test output, in the form of variables that have a Pearson correlation coefficient (r)> = 0.1 or 10%. These variables will be used to create multiple linear regression models to predict food supply based on variables from Facebook Audience Insights data. Table 4: Pearson Correlation Coefficient (r) with Domestic Supply Quantity 2014 Table 5: Pearson Correlation Coefficient (r) with Domestic Supply Quantity 2015 Table 6: Pearson Correlation Coefficient (r) with Domestic Supply Quantity 2016 Table 7: Pearson Correlation Coefficient (r) with Domestic Supply Quantity 2017 Table 8: Multiple linear regression statistical output, a more complete explanation for each column is in the draft paper.