In this study, a new signal model for shingled magnetic recording (SMR) is developed based on the drive waveform analysis with squeezed side tracks. Shingled recording effects are modeled by the data dependent effective write width or erase band, and the readback signals are represented by the weighted sum of the individual target and side track signals without intertrack interference (ITI). The weights of the target and side track signals, denoted by the ITI coefficients, are estimated from the drive waveforms collected from the track squeeze tests. Near the track edge location, the ITI coefficients are reduced by the erase band and the data-independent electronic noise is effectively increased. In addition, the ITI coefficients also depend on the recorded data patterns due to such effect as track width broadening by low frequency data patterns. Cross-track ITI coefficient profiles are modeled by Gaussian read sensitivity and rectangular magnetization functions and applied for various SMR conditions. Two data sets of 50% and 44% bit transition rates are investigated to take into account data dependent erase band. Based on the proposed signal model, the cross-track bit error rate (BER) performance of the SMR are evaluated with the readback signals, synthesized for various read offsets. The results show that modulating the data bits to have fewer bit transitions may improve the on-track BER performance, while its off-track performance can be limited by the increased ITI due to the low frequency data patterns. In addition, BER performance of SMR can be improved by the ITI cancellation (ITIC) scheme.