我们发现自己已经身处“大数据”时代之中。随着计算机的发展，我们拥有了在可穿戴设备，个人电脑，以至于云上存储，传输，计算以及分析“大数据”数据的技术能力，但是对于“大数据”概念本身，依然存在很多悬而未决的问题。在这个热门话题的背后，不同的描述“大数据”的方法，不同的解决“大数据”问题的方案往往将我们带去不同的方向，个中分析数据的过程，其研究范式也不尽相同。用两个案例研究，本文试图追溯“大数据”问题的一般研究方法。案例一分析了在几位印象派画家的画作中空间尺度上的统计特征，指出了在画作之中的幂律分布特性，并将其与之前文献中普遍讨论的自相似与色噪声研究中的幂律分布加以区分。在本案例中我们参考实物照片，与艺术画作进行对照，用同样的算法加以处理。并引用艺术史以及艺术理论，对艺术家特别是画家，在其艺术创作范式上与物理学家相似的研究范式进行分析，指出二者共同面对客观自然，通过抽象化以及原子化的认识去处理自然数据，处理后的结果既具有艺术价值，亦有充分的物理价值；案例二分析了在国际外汇市场中的经济时间序列数据，从时间尺度上使用非平稳非线性的分析方法，研究了在2011 年日本东部大地震这一自然灾害发生后，伴随发生的国际外汇市场“地震”中的“冲击波”。在时间序列上把金融数据分解为相振子，使国际市场的诸多货币在物理上对应成为一个复杂网络，以此使用物理上研究描述复杂网络问题的理论去解决经济问题，并成功地再现了当时外汇市场剧烈变动的“冲击波”。以比较成功的案例研究，本文进一步讨论“大数据”问题研究典型的研究范式，一方面提出了在中等“介观”尺度处理，而非直接考虑全部“大数据”的数据分析方法，追求在物理意义上认识“大数据”；另一方面充分考虑“大数据”的跨学科特性，追求从不同角度去研究“大数据”，特别以文献中古典音乐作品里幂律分布的发现为例，结合音乐理论本身，说明类似发现中很难解释的问题事实上来自于对音乐本身的忽视，比如音阶频率的指数分布，要求用比较研究的方法，借鉴不同学科的研究角度，全面考虑，全面认识“大数据”中的物理意义。

We find ourselves in the age of big-data, with the capabilities granted by the development ofmodern computing systems, wearable, personal computers, or in the cloud, in storing, transfering,computation, and analysis of data. However it is yet questionable ‘what is big-data’in a number of practical ways when different answers via different perspective may lead us tospecific paradigms in approaching the data we are interested in.In two case studies, the first on the scaling phenomena in artistic paintings in spatialpatterns, in comparision against the frequently referred to patterned noises and the feature ofself-similarity, and real world photos, we were trying to distinguish, both from statistical physicsand the viewpoint of fine art, the scaling found in the impressionists’ masterpieces from thefractal in noise, while reasoning the relationship of paintings with the representational photos,to show that in paradigm the artists share the perception of nature with that of scientists;the second concerning the financial time series in the global Foreign Exchange market when in2011, a devastating earthquake struck Japan, inflicting literally an earthquake in the financialmarket, and we could make use of numerical methods in data analysis to approach the timeseries data to exhibit the shock-wave of the earthquake in the Foreign Exchange market, whentalking about the market as a complex network, the physicists’ paradigm for the analysis ofan impulse on a nonstationary and nonlinear system could also be successful to some extentdepicting the features and behaviors in economics.With the case studies, we will further relate to a discussion on how big-data problems couldbe approached, when referring to the analysis of mesoscopic, rather than confronting directlythe ‘big’, data; and to carefully thread the clues interdisciplinary from multiple aspects, in seekfor common features or patterns in a comparative means for a comprehensive understanding ofthe system with big-data.