Abstract. Although Data Mining (DM) has been applied extensively in information systems and identified as a crucial tool for automatic data analysis and enterprise knowledge inference, still the practicability of DM has been little explored in the construction industry in particular. This study is conducted using data from actual practice in the customer service department of the target enterprise. Starting with data preparation, decision tree analysis and domain knowledge inference, practical verification is performed and previously unknown knowledge is discovered. Two practical barriers to enterprise information mining are found: the separation of information among various information systems and lack of key information attributes. Hence, there is a significant limitation on practical information data mining to generate new information, and only with a proper information integration one can benefit from the potential practical application of DM.

Residential district street lamps are important public safety facilities for night illumination. Residential district street lamps are numerous, consume a great deal of power, and are difficult to repair and maintain, as they require high costs in both time and labor to maintain individual luminance and a steady power supply to both the system and the lines. An automatic monitoring maintenance management system for street lamps is put forward in this study, which maintains changes in the loop currents of street lamp switch boxes, and uses the function of GPRS to send detected current change values to a control center in order to identify failures of street lamps of different loops, thus, both the cost of maintenance labor and maintenance time can be greatly reduced. The switching of street lamps can be controlled by a control center, and the utility management of energy and safety of the public can be effectively improved.