AIによる個人カーボンフットプリント・トラッカー
AI-Based Personal Carbon Footprint Tracker (原題)
Jogi John, Janhvi Mahakulkar, Harsh Sharma, Samiksha Dayare, Aditi Shende, Anjali Kushwaha
🤖 gxceed AI 要約
日本語
個人の日常活動を記録し、排出係数を用いてkg CO₂eを算定・集計するWebプラットフォームの設計と試作を提示。React/PHP/MySQL構成にGroq API経由のLLMを組み込み、カテゴリ別結果と個別助言を提供する。10名・14活動のパイロットでは合計112.51 kg CO₂e、消費カテゴリが53.3%を占め、手計算との一致も確認した。ただし小規模・短期・試算係数であり、予測やクラスタリングは未検証の拡張提案に留まる。
English
This paper presents the design and prototype of a web platform for personal carbon-footprint tracking, combining a React/PHP/MySQL stack with an LLM service via the Groq API. Users log activities, which are multiplied by stored emission factors and aggregated by category and period. A 10-user pilot over one month logged 14 activities totaling 112.51 kg CO₂e, with consumption at 53.3%; hand calculations matched all samples. Results are descriptive only, and proposed ML extensions remain unvalidated.
Unofficial AI-generated summary based on the public title and abstract. Not an official translation.
📝 gxceed 編集解説 — Why this matters
日本のGX文脈において
個人単位の排出可視化はScope 3下流・消費者向け脱炭素の接点として、SSBJや有報でのバリューチェーン開示が進む日本企業の顧客エンゲージメント設計に示唆を与える。ただし算定係数の検証とAI助言の品質評価が前提となる。
In the global GX context
Individual-level carbon tracking connects to downstream Scope 3 and consumer-facing decarbonization, an area global disclosure frameworks (CSRD, ISSB) increasingly scrutinize. The prototype's transparent activity-based calculation workflow offers a reference model, though verified emission factors and systematic LLM-advice evaluation are prerequisites for credible claims.
👥 読者別の含意
🔬研究者:AI×炭素会計の実装例として、排出係数ベース算定とLLM助言を組み合わせた評価設計の出発点になる。
🏢実務担当者:顧客・従業員向けの排出可視化ツール設計やScope 3下流エンゲージメントの参考になる。
🏛政策担当者:個人炭素管理の実装可能性を示すが、係数標準化とAI助言の品質保証が政策課題となる。
📄 Abstract（原文）
This paper presents the design and prototype of an intelligent web platform for personal carbon-footprint tracking. The platform enables users to record everyday activities, estimate associated greenhouse-gas emissions in kilograms of carbon-dioxide equivalent (kg CO₂e), review category-wise results, and receive personalised sustainability guidance. The architecture comprises a React, Tailwind CSS and Vite front end, a PHP back end, a MySQL database, and a large-language-model service accessed through the Groq API. The implemented calculation engine estimates emissions by multiplying each activity quantity by its stored emission factor and aggregates results by category and reporting period. A pilot dataset containing records from 10 users over September 1–30, 2026, included 14 logged activities and a reported total of 112.51 kg CO₂e. The mean across all 10 users was 11.25 kg CO₂e per user; the reported mean among active users was 22.50 kg CO₂e. The consumption category accounted for 53.3% of the reported total. A hand-calculation check matched the system output for all 14 samples. These findings are descriptive and limited by the small sample, short observation period, and use of illustrative emission factors. Random Forest forecasting, K-means user profiling, month-to-month comparisons, and leaderboard scoring are proposed extensions and have not been validated. The prototype demonstrates a transparent activity-based calculation workflow, while larger datasets, verified emission factors, and systematic evaluation of AI-generated advice are needed before broader claims can be made.
🔗 Provenance — このレコードを発見したソース
- openalex https://publications.ngmc.ac.in/journal/index.php/arjst/article/view/ai-based-personal-carbon-footprint-trackerfirst seen 2026-11-10 05:04:52 · last seen 2026-11-14 05:13:36
🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録（無料・3キーワードまで）。
gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。