Household air pollution is arising more and more concerns due to its severe impact on human health. In recent years, low cost sensors are more and more popular due to their advantages of data with high time resolution. However, the analysis methods for the large-scale data obtained from sensors are still rare, limiting the expression of the value of the high time resolution data. In this paper, we introduce two new indexes for the characterization of indoor PM_2.5 peak, aiming to evaluate the speed of the PM_2.5 increasing and decreasing (namely K_i and K_d, respectively) with internal emission sources starting and ending, and the reliability of the indexes was identified by using real world measurement. Large K_i is found in the residential homes during cooking time, indicating the urgent need to control PM_2.5 emissions from residential energy use and cooking oil. It is found that the K_i and K_d in homes burning solid fuels are higher than that using clean fuels, suggesting the potential benefit of energy transition in residential homes. This study highlighted the importance of various indexes used for indoor PM_2.5 peak characterization and it is expected to provide new insights for future indoor air pollution study.