In reinforced concrete (RC) structures, splicing is required owing to the length limitations of rebars, insufficient lengths, and transportation issues. In particular, splices connect the rebar of RC structures such as walls, columns, beams, slabs, and joints. Lap splices are the most commonly used method worldwide because they do not require specific equipment or skilled workers. However, lap splices incur high construction costs because of the long splice lengths required for large-diameter rebars in megastructures, as well as issues pertaining to material supply, labor costs, constructability, and project duration. Additionally, approximately 15% more rebar is required because of the overlap. Energy saving for a sustainable built environment is possible if the disadvantage of lap splices, which generate high CO2 emissions due to the excessive use of rebar, are resolved. Hence, mechanical rebar couplers (MRCs) have been developed. However, despite their advantages, they have not been widely applied in construction sites owing to concerns regarding safety, quality, and constructability. Although various MRCs have been developed, most studies focus only on their structural performance. Therefore, a data-driven approach for selecting MRCs based on the reinforcing bar shape and structural characteristics is proposed in this study. Using a data-driven MRC selection algorithm, using the T-threaded coupler for one rebar over two floors resulted in 56% more efficient labor productivity, 15% shorter assembly time, 17% lower costs, and 26% lower CO2 emission. Using a developed algorithm, the appropriate MRC can easily and rapidly be selected for frequent design changes.