Consumer Price Index Intern (P.I.C.K.)

Consumer Price Indices Multilateral Approach Study and Enhancement of Product Item Classification Kit (P.I.C.K.) National statistical offices are exploring multilateral index methods to modernise Consumer Price Index (CPI) compilation. Under this project, the intern will assist in developing an experimental multilateral CPI, including data cleaning, analysis, and evaluation. The project also involves enhancing P.I.C.K., an existing machine learning system that classifies web-scraped clothing prices into CPI product categories. The intern will contribute to developing web-scraping, Machine Learning (ML) algorithms, and Large Language Model (LLM) solutions to improve P.I.C.K. and extend it to additional CPI groups, supporting more efficient CPI compilation. Responsibilities a) Support the development of an experimental CPI series based on the multilateral index approach. i. developing python scripts and/or ML models to clean and classify scanner data into pre-defined expenditure categories ii. computation of multilateral indices, with reference to what other National Statistical Offices have done iii. analysis of the multilateral indices e.g., how it compares to the existing CPI b) Enhance the existing Machine Learning (ML) algorithm (P.I.C.K.) to improve the homogeneity of product classification and expand the use of enhanced version to other CPI groups, such as household durables. Requirements Knowledge or experience in data science, statistics, mathematics, machine learning, or related quantitative areas. Knowledge in open source tools such as Python, R, and experience working with Machine Learning (ML)/AI models would be beneficial. Internship Period: 6 months, from Jan to Jun 2027 (flexible start and end dates) Part-time option available: Minimum commitment of 3 days per week