Initial yields are used by institutional investors and investment managers to assess the pricing conditions of real estate markets. In contrast to commercial real estate, initial yields in the residential sector are hard to quantify, especially due to the lack of comparables. In the era of digitalisation and big data residential assets are mostly brought to the market via digital multiple listing systems. The paper develops semiparametric hedonic models for extracting the implicit information to calculate residential net initial yields for both a buy-to-hold and rental investment strategy based on more than 3 million observations. The results are robust and confirm that the pricing conditions of residential markets are captured by the hedonic approach, enhancing the transparency in real estate markets.