This project aims to develop a methodology to nowcast macroeconomic indicators such as GDP, retail trade sales and retail e-commerce sales with real-time data source, Google Trends. The volume of queries for different keywords and categories from Google Trends API are serving as the predictors for nowcasting the desired economic factors.
The key goals of the project are discussed below:
-
Nowcasting quarterly GDP: Our first goal is to nowcast the macroeconomic indicator GDP quarterly by using the real time Google Trends variables as predictors. The aim is to nowcast the GDP at national level and industry-wise.
-
Nowcasting monthly retail trade sales: The retail sales data are available monthly, so our objective is to nowcast the monthly retail trade sales industry-wise using real time data.
-
Nowcasting retail e-commerce sales: The retail sales data are available monthly, so our objective is to nowcast the monthly retail e-commerce sales industry-wise using real time data.
‘Nowcasting’ (or “predicting the present”) ; Choi and Varian
- Aishwarya Sharma: A data science enthusiast who loves to play with data and is always eager to explore more.
- Harpreet Kaur: Passionate mathematician on the path of becoming a statistician.
- Jagdeep Brar: A data science student who loves cooking and painting in free time!
The information on economic indicators is crucial for policy and decision making at an appropriate time. However, such information is usually available with a lag period. The nowcasting of economic indicators can help Statistics Canada (StatsCan) to fill such lags and make economy related decision without facing any delay in the information. The economic indicators are the representatives of economic activities of a country which may be explained by Google Trends. So, the goal of this project is to develop a methodology to nowcast macroeconomic indicators such as Gross Domestic Product (GDP), Retail Trade Sales (RTS) and Retail E-Commerce (EC) Sales in real time by using the real time data source, Google Trends. Google Trends provide daily, weekly, and monthly reports on the volume of Google queries related to different industries/keywords which can exhibit the business cycles and provide signals about the multiple aspects of the economy.
Macroeconomic factors are the key drivers of economy, and their timely information helps in good policy making. However, this information is available with a lag, for instance, the data for the present month’s GDP is generally published in the coming month/quarter which causes delay in decision-making. To overcome this issue of delayed information gave rise to nowcasting approach. This is what this project will serve. This project will nowcast the macroeconomic factors and will use the keywords from Google Trends to predict the indicators. A dashboard will be created which will be user friendly and will be interactive in nature.
Data set for this project are open ended and the short description about data is provided below:
-
Gross Domestic Product (GDP) at basic prices quarterly: This dataset is comma separated file containing the information about the GDP quarterly. This file contains data from 1997 and do have some missing values, thus will require data wrangling.
-
Retail trade sales by industry: This is a comma separated data set containing the information about the retail sales trades as per the industry. Data wrangling is required in this dataset as well.
-
Retail E-commerce sales: This is a comma separated dataset containing the information about the retail e-commerce sales. This is the data for digital sales and will require some data wrangling as well.
-
Google Trends API: We will be accessing Google Trends website to get real time data for the macroeconomic indicators. Different keywords, categories and subcategories will be used to extract Google Trends predictors such as Economic crisis, loans, GPS, unemployment, affordable housing, economy news, agriculture, and forestry.
Our focus will be on the data starting from 2004 as we have Google trends available from that period and this will provide us huge data for our nowcasting.
- H. Choi, H. Varian, Predicting the present with Google Trends, Economic record, 88 (2012), 2-9.
- Stock, J.H. and Watson, M.W., 2016. Dynamic factor models, factor-augmented vector autoregressions, and structural vector autoregressions in macroeconomics. In Handbook of macroeconomics (Vol. 2, pp. 415-525). Elsevier.
- Woloszko, N. (2020). Tracking activity in real time with Google Trends, OECD Economics Department Working Papers, No. 1634, OECD Publishing, Paris.
- Dauphin, M.J.F., Dybczak, M.K., Maneely, M., Sanjani, M.T., Suphaphiphat, M.N., Wang, Y. and Zhang, H., 2022. Nowcasting GDP-A Scalable Approach Using DFM, Machine Learning and Novel Data, Applied to European Economies. International Monetary Fund.
- Richardson, A., van Florenstein Mulder, T. and Vehbi, T., 2021. Nowcasting GDP using machine-learning algorithms: A real-time assessment. International Journal of Forecasting, 37(2), pp.941-948.