FORECASTING ELECTRIC CAR SALES BASED ON TIME SERIES ANALYSIS AND TWITTER (X) USER SENTIMENT

Ikhsan, Fikri Fadilah (2025) FORECASTING ELECTRIC CAR SALES BASED ON TIME SERIES ANALYSIS AND TWITTER (X) USER SENTIMENT. S1 thesis, Universitas Andalas.

	Text (Cover dan Abstrak) abstrak.pdf - Published Version Download (388kB)
	Text (Bab 1 Pendahuluan) Bab 1 Pendahuluan.pdf - Published Version Download (362kB)
	Text (Bab 5 Penutup) Bab 5 Penutup.pdf - Published Version Download (221kB)
	Text (Daftar Pustaka) Daftar Pustaka.pdf - Published Version Download (304kB)
	Text (Skripsi full text) skripsi fulltext.pdf - Published Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Electric cars are a means of land transportation that adopts electrical energy as the main source of power to run the driving motor. These vehicles are an integral part of the global effort to reduce greenhouse gas emissions and address climate change issues. However, suboptimal supporting ecosystems hinder the transition process of electric car adoption in Indonesia and substantially increase production uncertainty in the automotive industry sector. In response to this issue, various public sentiments can be observed through social media discussions and have an impact on consumer decisions before purchasing electric cars. The involvement of Twitter (X) social media in shaping public opinion is proposed as one of the significant factors influencing the sales of innovative products at the early stage of technology adoption. The findings of this study indicate that negative public sentiment toward electric cars in Indonesia primarily stems from three key issues: concerns regarding the inadequate battery waste management infrastructure, skepticism about the overall environmental impact of electric vehicles, and perceptions of inefficiency in government policy, particularly in the administration of subsidy programs aimed at accelerating EV adoption. The forecasting model that integrates negative sentiment variables into a regression framework yields the most optimal performance, with a Mean Absolute Error (MAE) of 1941.79 and a Weighted Mean Absolute Percentage Error (wMAPE) of 24%. These results suggest that widespread negative sentiment does not necessarily hinder electric vehicle adoption; rather, in certain contexts, it may increase public attention and positively influence purchasing decisions.

Actions (login required)

View Item